Tested configurations
What has actually been run, where, and how it went. The README’s status table sums this up by vendor; the runs below keep the evidence.
Read the combined catalogue on the website for the frozen table and new run files together. This source page keeps the existing table visible without running a command. The website assembles its view from these repository files; it does not collect results from a service.
Browse the new run files individually, or use the combined catalogue above. For each new measured run,
add a self-contained YYYY-MM-DD-topic-tester-run.md file under docs/tested/, following the
format and checking instructions. The tester’s name and a distinct run identifier are required;
repeat the context instead of saying as above. New results belong in those files, not in the table below.
The website orders new files by date, then filename, descending. Logs and measured commits still belong in the pull request’s Measurements and Testing sections. Do not commit an assembled reading view.
New runs
| Date | Build | OS / desktop | GPU / driver | CPU / RAM | Settings | Where | Result |
|---|---|---|---|---|---|---|---|
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and #446’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | RX_VK_UB_SNAPSHOT=0 RX_VK_PUSH_ALL=1 (the uniform snapshot off, every descriptor pushed at every draw; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the village (child), then the Market (child) in one session by the debug teleport: 40 s to settle at each arrival spot, then 40 s walking in each (the left stick, repeated: forward 150 frames, right 20, forward 90, left 45); fps from the log’s 5 s vi: windows, dark and bright patches from the region test on every presented frame of the measured windows (2 or more of 16x9 regions more than 10 levels darker, or brighter, than in both neighbouring frames); one of 4 rounds of three sessions in turning orders: XR1 XA1 XP1, XP2 XA2 XR2, XA3 XP3 XR3, XR4 XP4 XA4 (R #446 alone, A with #952, P the snapshot off with every descriptor pushed); then walks YA2 YP2 YP3 YA3 YA4 YP4 |
the village 65.8 and the Market 53.4 fps walking (medians of 8 and 8 5 s windows; 940 and 1,384 draws a frame); by the region test’s last line, 13 of 4,800 frames compared over both windows with a dark patch, 9 with a bright one (the largest darkening 41.4); whole pictures dark 0 in the village and 0 in the Market; no device loss (run YP4, the snapshot off, every descriptor pushed) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and #446’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | RX_VK_UB_SNAPSHOT=0 RX_VK_PUSH_ALL=1 (the uniform snapshot off, every descriptor pushed at every draw; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the village (child), then the Market (child) in one session by the debug teleport: 40 s to settle at each arrival spot, then 40 s walking in each (the left stick, repeated: forward 150 frames, right 20, forward 90, left 45); fps from the log’s 5 s vi: windows, dark and bright patches from the region test on every presented frame of the measured windows (2 or more of 16x9 regions more than 10 levels darker, or brighter, than in both neighbouring frames); one of 4 rounds of three sessions in turning orders: XR1 XA1 XP1, XP2 XA2 XR2, XA3 XP3 XR3, XR4 XP4 XA4 (R #446 alone, A with #952, P the snapshot off with every descriptor pushed); then walks YA2 YP2 YP3 YA3 YA4 YP4 |
the village 64.9 and the Market 63.9 fps walking (medians of 8 and 9 5 s windows; 1,404 and 1,513 draws a frame); by the region test’s last line, 8 of 5,400 frames compared over both windows with a dark patch, 13 with a bright one (the largest darkening 39.2); whole pictures dark 0 in the village and 0 in the Market; no device loss (run YP3, the snapshot off, every descriptor pushed) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and #446’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | RX_VK_UB_SNAPSHOT=0 RX_VK_PUSH_ALL=1 (the uniform snapshot off, every descriptor pushed at every draw; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the village (child), then the Market (child) in one session by the debug teleport: 40 s to settle at each arrival spot, then 40 s walking in each (the left stick, repeated: forward 150 frames, right 20, forward 90, left 45); fps from the log’s 5 s vi: windows, dark and bright patches from the region test on every presented frame of the measured windows (2 or more of 16x9 regions more than 10 levels darker, or brighter, than in both neighbouring frames); one of 4 rounds of three sessions in turning orders: XR1 XA1 XP1, XP2 XA2 XR2, XA3 XP3 XR3, XR4 XP4 XA4 (R #446 alone, A with #952, P the snapshot off with every descriptor pushed); then walks YA2 YP2 YP3 YA3 YA4 YP4 |
the village 64.7 and the Market 53.5 fps walking (medians of 9 and 9 5 s windows; 656 and 1,145 draws a frame); by the region test’s last line, 3 of 4,800 frames compared over both windows with a dark patch, 10 with a bright one (the largest darkening 38.0); whole pictures dark 0 in the village and 0 in the Market; no device loss (run YP2, the snapshot off, every descriptor pushed) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and #446’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | RX_VK_UB_SNAPSHOT=0 RX_VK_PUSH_ALL=1 (the uniform snapshot off, every descriptor pushed at every draw; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 30 s measured in each; fps from the log’s 5 s vi: windows, dark and bright patches from the region test on every presented frame of the measured windows (2 or more of 16x9 regions more than 10 levels darker, or brighter, than in both neighbouring frames); one of 4 rounds of three sessions in turning orders: XR1 XA1 XP1, XP2 XA2 XR2, XA3 XP3 XR3, XR4 XP4 XA4 (R #446 alone, A with #952, P the snapshot off with every descriptor pushed); then walks YA2 YP2 YP3 YA3 YA4 YP4 |
the starting forest 89.1 and the Market 57.5 fps standing (medians of 7 and 6 5 s windows; 808 and 2,488 draws a frame); by the region test’s last line, 0 of 4,200 frames compared over both windows with a dark patch, 0 with a bright one (the largest darkening 0.7); whole pictures dark 0 in the starting forest and 0 in the Market; no device loss (run XP4, the snapshot off, every descriptor pushed) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and #446’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | RX_VK_UB_SNAPSHOT=0 RX_VK_PUSH_ALL=1 (the uniform snapshot off, every descriptor pushed at every draw; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 30 s measured in each; fps from the log’s 5 s vi: windows, dark and bright patches from the region test on every presented frame of the measured windows (2 or more of 16x9 regions more than 10 levels darker, or brighter, than in both neighbouring frames); one of 4 rounds of three sessions in turning orders: XR1 XA1 XP1, XP2 XA2 XR2, XA3 XP3 XR3, XR4 XP4 XA4 (R #446 alone, A with #952, P the snapshot off with every descriptor pushed); then walks YA2 YP2 YP3 YA3 YA4 YP4 |
the starting forest 90.2 and the Market 53.5 fps standing (medians of 6 and 6 5 s windows; 782 and 2,548 draws a frame); by the region test’s last line, 0 of 4,200 frames compared over both windows with a dark patch, 0 with a bright one (the largest darkening 2.3); whole pictures dark 0 in the starting forest and 0 in the Market; no device loss (run XP3, the snapshot off, every descriptor pushed) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and #446’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | RX_VK_UB_SNAPSHOT=0 RX_VK_PUSH_ALL=1 (the uniform snapshot off, every descriptor pushed at every draw; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 30 s measured in each; fps from the log’s 5 s vi: windows, dark and bright patches from the region test on every presented frame of the measured windows (2 or more of 16x9 regions more than 10 levels darker, or brighter, than in both neighbouring frames); one of 4 rounds of three sessions in turning orders: XR1 XA1 XP1, XP2 XA2 XR2, XA3 XP3 XR3, XR4 XP4 XA4 (R #446 alone, A with #952, P the snapshot off with every descriptor pushed); then walks YA2 YP2 YP3 YA3 YA4 YP4 |
the starting forest 90.5 and the Market 57.6 fps standing (medians of 6 and 6 5 s windows; 787 and 2,464 draws a frame); by the region test’s last line, 0 of 4,200 frames compared over both windows with a dark patch, 0 with a bright one (the largest darkening 0.6); whole pictures dark 0 in the starting forest and 0 in the Market; no device loss (run XP2, the snapshot off, every descriptor pushed) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and #446’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | RX_VK_UB_SNAPSHOT=0 RX_VK_PUSH_ALL=1 (the uniform snapshot off, every descriptor pushed at every draw; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 30 s measured in each; fps from the log’s 5 s vi: windows, dark and bright patches from the region test on every presented frame of the measured windows (2 or more of 16x9 regions more than 10 levels darker, or brighter, than in both neighbouring frames); one of 4 rounds of three sessions in turning orders: XR1 XA1 XP1, XP2 XA2 XR2, XA3 XP3 XR3, XR4 XP4 XA4 (R #446 alone, A with #952, P the snapshot off with every descriptor pushed); then walks YA2 YP2 YP3 YA3 YA4 YP4 |
the starting forest 91.8 and the Market 55.1 fps standing (medians of 6 and 6 5 s windows; 792 and 2,694 draws a frame); by the region test’s last line, 0 of 4,200 frames compared over both windows with a dark patch, 0 with a bright one (the largest darkening 0.7); whole pictures dark 0 in the starting forest and 0 in the Market; no device loss (run XP1, the snapshot off, every descriptor pushed) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and #446’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on, as on every Intel GPU with #446; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 30 s measured in each; fps from the log’s 5 s vi: windows, dark and bright patches from the region test on every presented frame of the measured windows (2 or more of 16x9 regions more than 10 levels darker, or brighter, than in both neighbouring frames); one of 4 rounds of three sessions in turning orders: XR1 XA1 XP1, XP2 XA2 XR2, XA3 XP3 XR3, XR4 XP4 XA4 (R #446 alone, A with #952, P the snapshot off with every descriptor pushed); then walks YA2 YP2 YP3 YA3 YA4 YP4 |
the starting forest 86.5 and the Market 53.4 fps standing (medians of 7 and 6 5 s windows; 740 and 2,460 draws a frame); by the region test’s last line, 0 of 4,200 frames compared over both windows with a dark patch, 0 with a bright one (the largest darkening 1.8); whole pictures dark 0 in the starting forest and 0 in the Market; no device loss (run XR4, #446 alone) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and #446’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on, as on every Intel GPU with #446; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 30 s measured in each; fps from the log’s 5 s vi: windows, dark and bright patches from the region test on every presented frame of the measured windows (2 or more of 16x9 regions more than 10 levels darker, or brighter, than in both neighbouring frames); one of 4 rounds of three sessions in turning orders: XR1 XA1 XP1, XP2 XA2 XR2, XA3 XP3 XR3, XR4 XP4 XA4 (R #446 alone, A with #952, P the snapshot off with every descriptor pushed); then walks YA2 YP2 YP3 YA3 YA4 YP4 |
the starting forest 88.6 and the Market 54.2 fps standing (medians of 6 and 6 5 s windows; 794 and 2,712 draws a frame); by the region test’s last line, 0 of 4,200 frames compared over both windows with a dark patch, 0 with a bright one (the largest darkening 0.5); whole pictures dark 0 in the starting forest and 0 in the Market; no device loss (run XR3, #446 alone) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and #446’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on, as on every Intel GPU with #446; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 30 s measured in each; fps from the log’s 5 s vi: windows, dark and bright patches from the region test on every presented frame of the measured windows (2 or more of 16x9 regions more than 10 levels darker, or brighter, than in both neighbouring frames); one of 4 rounds of three sessions in turning orders: XR1 XA1 XP1, XP2 XA2 XR2, XA3 XP3 XR3, XR4 XP4 XA4 (R #446 alone, A with #952, P the snapshot off with every descriptor pushed); then walks YA2 YP2 YP3 YA3 YA4 YP4 |
the starting forest 86.9 and the Market 54.5 fps standing (medians of 6 and 6 5 s windows; 778 and 2,576 draws a frame); by the region test’s last line, 0 of 4,200 frames compared over both windows with a dark patch, 0 with a bright one (the largest darkening 1.1); whole pictures dark 0 in the starting forest and 0 in the Market; no device loss (run XR2, #446 alone) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and #446’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on, as on every Intel GPU with #446; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 30 s measured in each; fps from the log’s 5 s vi: windows, dark and bright patches from the region test on every presented frame of the measured windows (2 or more of 16x9 regions more than 10 levels darker, or brighter, than in both neighbouring frames); one of 4 rounds of three sessions in turning orders: XR1 XA1 XP1, XP2 XA2 XR2, XA3 XP3 XR3, XR4 XP4 XA4 (R #446 alone, A with #952, P the snapshot off with every descriptor pushed); then walks YA2 YP2 YP3 YA3 YA4 YP4 |
the starting forest 86.8 and the Market 54.6 fps standing (medians of 6 and 6 5 s windows; 793 and 2,567 draws a frame); by the region test’s last line, 0 of 4,200 frames compared over both windows with a dark patch, 0 with a bright one (the largest darkening 0.7); whole pictures dark 0 in the starting forest and 0 in the Market; no device loss (run XR1, #446 alone) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and this PR’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on, as on every Intel GPU with this PR; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the village (child), then the Market (child) in one session by the debug teleport: 40 s to settle at each arrival spot, then 40 s walking in each (the left stick, repeated: forward 150 frames, right 20, forward 90, left 45); fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames), the guard’s counts from the log’s copied into the stream line |
the village 70.5 and the Market 56.9 fps walking (medians of 9 and 8 5 s windows; 1,595 and 1,085 draws a frame); 2 of 4,800 frames over both windows with a dark patch (the largest change 37.2), 5 with a bright one; whole pictures dark 0 in the village and 0 in the Market; by the session’s last count the guard bound no uniform range in place: none after a GPU write in its submission, none after one in a submission still on the GPU; no device loss (run WX2) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and this PR’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on, as on every Intel GPU with this PR; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the village (child), then the Market (child) in one session by the debug teleport: 40 s to settle at each arrival spot, then 40 s walking in each (the left stick, repeated: forward 150 frames, right 20, forward 90, left 45); fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames), the guard’s counts from the log’s copied into the stream line |
the village 60.7 and the Market 61.8 fps walking (medians of 8 and 9 5 s windows; 959 and 1,538 draws a frame); 7 of 4,800 frames over both windows with a dark patch (the largest change 48.2), 17 with a bright one; whole pictures dark 0 in the village and 0 in the Market; by the session’s last count the guard bound no uniform range in place: none after a GPU write in its submission, none after one in a submission still on the GPU; no device loss (run WX1) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0); without this PR’s guard |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on, as on every Intel GPU with this PR; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 40 s measured in each; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames), the guard’s counts from the log’s copied into the stream line |
the starting forest 88.0 and the Market 53.4 fps (medians of 8 and 8 5 s windows; 790 and 2,591 draws a frame); 0 of 5,400 frames over both windows with a dark patch (the largest change 0.7); no device loss (run XU4) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and this PR’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on, as on every Intel GPU with this PR; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 40 s measured in each; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames), the guard’s counts from the log’s copied into the stream line |
the starting forest 88.3 and the Market 54.2 fps (medians of 8 and 8 5 s windows; 801 and 2,652 draws a frame); 0 of 5,400 frames over both windows with a dark patch (the largest change 0.9); by the session’s last count the guard bound no uniform range in place: none after a GPU write in its submission, none after one in a submission still on the GPU; no device loss (run UX4) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and this PR’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on, as on every Intel GPU with this PR; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 40 s measured in each; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames), the guard’s counts from the log’s copied into the stream line |
the starting forest 87.7 and the Market 54.0 fps (medians of 8 and 8 5 s windows; 782 and 2,557 draws a frame); 0 of 5,400 frames over both windows with a dark patch (the largest change 1.2); by the session’s last count the guard bound no uniform range in place: none after a GPU write in its submission, none after one in a submission still on the GPU; no device loss (run UX3) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0); without this PR’s guard |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on, as on every Intel GPU with this PR; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 40 s measured in each; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames), the guard’s counts from the log’s copied into the stream line |
the starting forest 87.9 and the Market 52.7 fps (medians of 8 and 9 5 s windows; 776 and 2,432 draws a frame); 0 of 5,400 frames over both windows with a dark patch (the largest change 0.6); no device loss (run XU3) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and this PR’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on, as on every Intel GPU with this PR; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 40 s measured in each; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames), the guard’s counts from the log’s copied into the stream line |
the starting forest 87.8 and the Market 55.3 fps (medians of 9 and 8 5 s windows; 816 and 2,504 draws a frame); 0 of 5,400 frames over both windows with a dark patch (the largest change 1.0); by the session’s last count the guard bound no uniform range in place: none after a GPU write in its submission, none after one in a submission still on the GPU; no device loss (run UX2) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0); without this PR’s guard |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on, as on every Intel GPU with this PR; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 40 s measured in each; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames), the guard’s counts from the log’s copied into the stream line |
the starting forest 87.1 and the Market 55.4 fps (medians of 8 and 8 5 s windows; 780 and 2,674 draws a frame); 0 of 5,400 frames over both windows with a dark patch (the largest change 1.3); no device loss (run XU2) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0); without this PR’s guard |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on, as on every Intel GPU with this PR; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 40 s measured in each; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames), the guard’s counts from the log’s copied into the stream line |
the starting forest 88.6 and the Market 55.1 fps (medians of 8 and 8 5 s windows; 747 and 2,427 draws a frame); 0 of 5,400 frames over both windows with a dark patch (the largest change 1.7); no device loss (run XU1) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and this PR’s guard: the renderer diff of 32ab00b5..c10f1e32 (fix/arc-discrete-ub-snapshot at c10f1e32), as a local commit on that source |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on, as on every Intel GPU with this PR; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 40 s measured in each; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames), the guard’s counts from the log’s copied into the stream line |
the starting forest 86.9 and the Market 52.0 fps (medians of 9 and 8 5 s windows; 767 and 2,480 draws a frame); 0 of 5,400 frames over both windows with a dark patch (the largest change 0.9); by the session’s last count the guard bound no uniform range in place: none after a GPU write in its submission, none after one in a submission still on the GPU; no device loss (run UX1) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), this PR’s guard as it was before (the renderer diff of 32ab00b5..0f0d82af: bound in place within a submission only), and a local timing (in no pull request): of the guard’s checks in uniform_range, one in three timed whole in TSC ticks (3.5 GHz), one in three an empty timed region, one in three in parts; each note and each submission’s start timed |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a timed job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 40 s measured in each; the timing’s log line every 2^20 checks, its last line in each window; fps from the log’s 5 s vi: windows |
the starting forest 85.3 and the Market 52.3 fps (medians of 9 and 8 5 s windows; 772 and 2,617 draws a frame), with the timing; the starting forest: 324,051 checks a second at 49.2 TSC ticks (14.1 ns) over the empty region, 23,676 notes a second at 604.3 ticks, 680 submissions a second (its map 12.6 ranges at their end on average, 103 at most; 2818 ticks to begin again), the guard 9.19 ms of thread time a second in all; the Market: 545,552 checks a second at 38.8 TSC ticks (11.1 ns) over the empty region, 16,562 notes a second at 540.7 ticks, 419 submissions a second (its map 16.2 ranges at their end on average, 134 at most; 4537 ticks to begin again), the guard 9.15 ms of thread time a second in all; no device loss (run GT1) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), this PR’s guard as it was before (the renderer diff of 32ab00b5..0f0d82af: bound in place within a submission only), and a local probe (in no pull request): at each uniform snapshot, under a lock, whether the range met a page an earlier submission still wrote back, a range an earlier submission’s GPU wrote (kept from the session’s start), and whether that range’s pages were ever written back |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a shared job of the queue |
the village (child), then the Market (child) in one session by the debug teleport: 40 s to settle at each arrival spot, then 40 s walking in each (the left stick, repeated: forward 150 frames, right 20, forward 90, left 45); the probe’s counts from its log line, fps from the log’s 5 s vi: windows |
the village 58.6 and the Market 56.8 fps walking (medians of 8 and 9 5 s windows; 1,495 and 2,331 draws a frame), with the probe’s lock; by the end of the village’s window 47,880,000 snapshots: 439,760 met a page an earlier submission still wrote back, 198,705 a range an earlier submission’s GPU wrote (198,705 written back since, 0 never), 172,595 both; by the end of the Market’s window 87,840,000 snapshots: 683,313 met a page an earlier submission still wrote back, 529,212 a range an earlier submission’s GPU wrote (529,212 written back since, 0 never), 378,540 both; no device loss (run PB2) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 (main 2b1d38c6 with #443, #446’s first commit, #459 and #500 merged) with the render-thread changes since merged to main (#761, #838, 64ed6a2f), #1083’s three commits and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), this PR’s guard as it was before (the renderer diff of 32ab00b5..0f0d82af: bound in place within a submission only), and a local probe (in no pull request): at each uniform snapshot, under a lock, whether the range met a page an earlier submission still wrote back, a range an earlier submission’s GPU wrote (kept from the session’s start), and whether that range’s pages were ever written back |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | the defaults (the uniform snapshot on; whole frames on); HOS_PROFILE=5, HOS_FPS_MAX=240; 1280x720 window, the scene rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (the Upscaler row’s Intel XeSS, Ultra Quality), frame generation off; a shared job of the queue |
the starting forest (child), then the Market (child) in one session by the debug teleport, standing at each arrival spot: 40 s to settle, then 40 s measured in each; the probe’s counts from its log line, fps from the log’s 5 s vi: windows |
the starting forest 85.4 and the Market 52.2 fps (medians of 8 and 8 5 s windows; 723 and 2,629 draws a frame), with the probe’s lock; by the end of the starting forest’s window 31,800,000 snapshots: 212,910 met a page an earlier submission still wrote back, 113,173 a range an earlier submission’s GPU wrote (113,173 written back since, 0 never), 67,750 both; by the end of the Market’s window 76,220,000 snapshots: 439,941 met a page an earlier submission still wrote back, 438,322 a range an earlier submission’s GPU wrote (438,322 written back since, 0 never), 256,053 both; no device loss (run PB1) |
| 2026-11-16 | longshot v3.0.0-rc.4-119-gd9486a08-dirty (source 624149ae): main d9486a08 with this change in the working tree (not committed when run); both sides the same build; built with make_game.py |
Windows 11 Pro 26200, no window (headless, scripted input) | RTX 4070 SUPER 12 GB, NVIDIA 617.42 | Ryzen 9 9950X3D (16 cores, SMT off), 32 GB | 2560x1440, DLAA, no frame generation, uncapped (HOS_FPS_MAX=1000); RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32, RX_VK_LOAD=create, RX_VK_MIRROR_SPARSE=1, RX_VK_TEST_ALLOWED_MB=10200; the render thread (Auto) |
Castle Town’s market at night: save set “First Step into the game world Town”, night at 40 s (!cheat 121), the walk into the market and the camera turning (town-market.txt, wall clock, 100 s); one round, the two sides in turn |
run pp-main-1, RX_VK_TARGET_POOL_MB=256 RX_VK_DEAD_EVERY=10 RX_VK_DEAD_MIN_UNUSED=3 (main’s behaviour): the market’s fps median 69.7 (the last 12 samples of 120 presents with 1,500-2,999 draws); render target images 25820 reused from the pool, 21501 made new, 21229 destroyed; frames over 25 ms 76 and over 50 ms 14 in the run’s last two thirds; 1 % low 42.1 fps |
| 2026-11-16 | longshot v3.0.0-rc.4-119-gd9486a08-dirty (source 624149ae): main d9486a08 with this change in the working tree (not committed when run); both sides the same build; built with make_game.py |
Windows 11 Pro 26200, no window (headless, scripted input) | RTX 4070 SUPER 12 GB, NVIDIA 617.42 | Ryzen 9 9950X3D (16 cores, SMT off), 32 GB | 2560x1440, DLAA, no frame generation, uncapped (HOS_FPS_MAX=1000); RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32, RX_VK_LOAD=create, RX_VK_MIRROR_SPARSE=1, RX_VK_TEST_ALLOWED_MB=10200; the render thread (Auto) |
Castle Town’s market at night: save set “First Step into the game world Town”, night at 40 s (!cheat 121), the walk into the market and the camera turning (town-market.txt, wall clock, 100 s); one round, the two sides in turn |
run pp-new-1, this change’s defaults (768 MB pool on this PC, dead targets every present after one unused): the market’s fps median 70.7 (the last 12 samples of 120 presents with 1,500-2,999 draws); render target images 34433 reused from the pool, 8096 made new, 7768 destroyed; frames over 25 ms 49 and over 50 ms 10 in the run’s last two thirds; 1 % low 44.8 fps |
| 2026-11-16 UTC | branch mma-unroll (sheep’s BunnyHood commit 01f9b690d, a cherry-pick of their e95c60a, cherry-picked onto Longshot main): its gpu/ change (head f1365aaf; emit_spirv_warp.cpp, mma_layout.h, mma_software.h and vk_coopmat.cpp byte-identical on the rebased head, the vk_shader.cpp stamp hunk the same lines) applied on main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed) as a renderer-only pack via scripts/dev_packs.py (longshot_vk.so sha256 28bddba5f5a1b7c307f299ec170dc5ea8de75ad1b84d747920281c12998e950e, the game binary main’s); the same pack in every run of both arms |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2-2-neptune-72), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), stock automatic GPU clock | Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared (MemTotal 14.45 GiB) | native: 1280x800 output, render resolution 100% (options output_res=1280x800, resolution=1, upscaling=2), the game’s own upscaler, frame generation off; no longshot_env.txt; empty program and pipeline caches in a fresh private game folder (the translator changed), programs translated as first drawn; switches: none besides RX_VK_PRECOMPILE=0 (the default: the unrolled form tried on RADV) |
the starting forest (child, the save’s spot outside the player character’s house beside Saria, about 2,200 draws a frame), 75 s warm-up then 45 s standing; one session 2026-11-16 04:54-05:46 UTC, timed runs interleaved loop, unrolled, unrolled, loop, then pass timing loop, unrolled, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy and APU power from sysfs at 10 Hz |
run mu-U-kN-2, started 05:29:53 UTC, arm unrolled (default): log once vk: matrix multiplies in software (no cooperative matrices) and vk: software matrix multiplies unrolled, f16 rounding by conversion, A through quad broadcasts: the loop's words for 8192 lanes (RX_VK_MMA_UNROLL=0: looped). 9.67 fps, 1% low 6.07 fps; present interval p50 / p95 / p99 98.1 / 141.3 / 160.5 ms, 193 over 100 ms; GPU time 72.65 ms a frame; GPU busy 85%, 19% of samples at 200 MHz, the rest mostly 1040 MHz; APU power mean / max 14.97 / 19.24 W; temperature max 69 C; lowest available memory 2233 MiB. Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | branch mma-unroll (sheep’s BunnyHood commit 01f9b690d, a cherry-pick of their e95c60a, cherry-picked onto Longshot main): its gpu/ change (head f1365aaf; emit_spirv_warp.cpp, mma_layout.h, mma_software.h and vk_coopmat.cpp byte-identical on the rebased head, the vk_shader.cpp stamp hunk the same lines) applied on main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed) as a renderer-only pack via scripts/dev_packs.py (longshot_vk.so sha256 28bddba5f5a1b7c307f299ec170dc5ea8de75ad1b84d747920281c12998e950e, the game binary main’s); the same pack in every run of both arms |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2-2-neptune-72), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), stock automatic GPU clock | Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared (MemTotal 14.45 GiB) | native: 1280x800 output, render resolution 100% (options output_res=1280x800, resolution=1, upscaling=2), the game’s own upscaler, frame generation off; no longshot_env.txt; empty program and pipeline caches in a fresh private game folder (the translator changed), programs translated as first drawn; switches: none besides RX_VK_PRECOMPILE=0 (the default: the unrolled form tried on RADV) |
the starting forest (child, the save’s spot outside the player character’s house beside Saria, about 2,200 draws a frame), 75 s warm-up then 45 s standing; one session 2026-11-16 04:54-05:46 UTC, timed runs interleaved loop, unrolled, unrolled, loop, then pass timing loop, unrolled, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy and APU power from sysfs at 10 Hz |
run mu-U-kN-1, started 05:25:48 UTC, arm unrolled (default): log once vk: matrix multiplies in software (no cooperative matrices) and vk: software matrix multiplies unrolled, f16 rounding by conversion, A through quad broadcasts: the loop's words for 8192 lanes (RX_VK_MMA_UNROLL=0: looped). 9.86 fps, 1% low 6.57 fps; present interval p50 / p95 / p99 97.2 / 133.8 / 146.5 ms, 175 over 100 ms; GPU time 69.35 ms a frame; GPU busy 79%, 22% of samples at 200 MHz, the rest mostly 1040 MHz; APU power mean / max 14.94 / 18.16 W; temperature max 69 C; lowest available memory 1985 MiB. Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | branch mma-unroll (sheep’s BunnyHood commit 01f9b690d, a cherry-pick of their e95c60a, cherry-picked onto Longshot main): its gpu/ change (head f1365aaf; emit_spirv_warp.cpp, mma_layout.h, mma_software.h and vk_coopmat.cpp byte-identical on the rebased head, the vk_shader.cpp stamp hunk the same lines) applied on main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed) as a renderer-only pack via scripts/dev_packs.py (longshot_vk.so sha256 28bddba5f5a1b7c307f299ec170dc5ea8de75ad1b84d747920281c12998e950e, the game binary main’s); the same pack in every run of both arms |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2-2-neptune-72), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), stock automatic GPU clock | Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared (MemTotal 14.45 GiB) | native: 1280x800 output, render resolution 100% (options output_res=1280x800, resolution=1, upscaling=2), the game’s own upscaler, frame generation off; no longshot_env.txt; empty program and pipeline caches in a fresh private game folder (the translator changed), programs translated as first drawn; switches: none besides RX_VK_PRECOMPILE=0 (the default: the unrolled form tried on RADV) |
the main outdoor area (child, teleported from the starting-area save), 75 s warm-up then 45 s standing; one session 2026-11-16 04:54-05:46 UTC, timed runs interleaved loop, unrolled, unrolled, loop, then pass timing loop, unrolled, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy and APU power from sysfs at 10 Hz |
run mu-U-fN-2, started 05:03:54 UTC, arm unrolled (default): log once vk: matrix multiplies in software (no cooperative matrices) and vk: software matrix multiplies unrolled, f16 rounding by conversion, A through quad broadcasts: the loop's words for 8192 lanes (RX_VK_MMA_UNROLL=0: looped). 16.00 fps, 1% low 11.70 fps; present interval p50 / p95 / p99 62.0 / 67.1 / 72.3 ms, 2 over 100 ms; GPU time 40.2 ms a frame; GPU busy 72%, 15% of samples at 200 MHz, the rest mostly 1040 MHz; APU power mean / max 8.02 / 10.25 W; temperature max 62 C; lowest available memory 1588 MiB. Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | branch mma-unroll (sheep’s BunnyHood commit 01f9b690d, a cherry-pick of their e95c60a, cherry-picked onto Longshot main): its gpu/ change (head f1365aaf; emit_spirv_warp.cpp, mma_layout.h, mma_software.h and vk_coopmat.cpp byte-identical on the rebased head, the vk_shader.cpp stamp hunk the same lines) applied on main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed) as a renderer-only pack via scripts/dev_packs.py (longshot_vk.so sha256 28bddba5f5a1b7c307f299ec170dc5ea8de75ad1b84d747920281c12998e950e, the game binary main’s); the same pack in every run of both arms |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2-2-neptune-72), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), stock automatic GPU clock | Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared (MemTotal 14.45 GiB) | native: 1280x800 output, render resolution 100% (options output_res=1280x800, resolution=1, upscaling=2), the game’s own upscaler, frame generation off; no longshot_env.txt; empty program and pipeline caches in a fresh private game folder (the translator changed), programs translated as first drawn; switches: none besides RX_VK_PRECOMPILE=0 (the default: the unrolled form tried on RADV) |
the main outdoor area (child, teleported from the starting-area save), 75 s warm-up then 45 s standing; one session 2026-11-16 04:54-05:46 UTC, timed runs interleaved loop, unrolled, unrolled, loop, then pass timing loop, unrolled, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy and APU power from sysfs at 10 Hz |
run mu-U-fN-1, started 04:59:22 UTC, arm unrolled (default): log once vk: matrix multiplies in software (no cooperative matrices) and vk: software matrix multiplies unrolled, f16 rounding by conversion, A through quad broadcasts: the loop's words for 8192 lanes (RX_VK_MMA_UNROLL=0: looped). 16.12 fps, 1% low 13.12 fps; present interval p50 / p95 / p99 61.8 / 66.7 / 69.6 ms, 1 over 100 ms; GPU time 39.9 ms a frame; GPU busy 74%, 12% of samples at 200 MHz, the rest mostly 1040 MHz; APU power mean / max 7.95 / 9.82 W; temperature max 62 C; lowest available memory 1455 MiB. Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | branch mma-unroll (sheep’s BunnyHood commit 01f9b690d, a cherry-pick of their e95c60a, cherry-picked onto Longshot main): its gpu/ change (head f1365aaf; emit_spirv_warp.cpp, mma_layout.h, mma_software.h and vk_coopmat.cpp byte-identical on the rebased head, the vk_shader.cpp stamp hunk the same lines) applied on main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed) as a renderer-only pack via scripts/dev_packs.py (longshot_vk.so sha256 28bddba5f5a1b7c307f299ec170dc5ea8de75ad1b84d747920281c12998e950e, the game binary main’s); the same pack in every run of both arms |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2-2-neptune-72), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), stock automatic GPU clock | Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared (MemTotal 14.45 GiB) | native: 1280x800 output, render resolution 100% (options output_res=1280x800, resolution=1, upscaling=2), the game’s own upscaler, frame generation off; no longshot_env.txt; empty program and pipeline caches in a fresh private game folder (the translator changed), programs translated as first drawn; switches: none besides RX_VK_PRECOMPILE=0 (the default: the unrolled form tried on RADV); RX_VK_PASS_TIMING=2 (profiled: absolute times inflated) |
the starting forest (child, the save’s spot outside the player character’s house beside Saria, about 2,200 draws a frame), 75 s warm-up then 45 s standing; one session 2026-11-16 04:54-05:46 UTC, timed runs interleaved loop, unrolled, unrolled, loop, then pass timing loop, unrolled, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy and APU power from sysfs at 10 Hz |
run mu-pt-U-kN, started 05:42:09 UTC, arm unrolled (default): log once vk: matrix multiplies in software (no cooperative matrices) and vk: software matrix multiplies unrolled, f16 rounding by conversion, A through quad broadcasts: the loop's words for 8192 lanes (RX_VK_MMA_UNROLL=0: looped). Profiled: 9.46 fps, 1% low 6.02 fps; present interval p50 / p95 / p99 99.2 / 144.8 / 160.9 ms, 210 over 100 ms; GPU time 73.0 ms a frame; GPU busy 82%, 25% of samples at 200 MHz, the rest mostly 1040 MHz; APU power mean / max 14.92 / 18.32 W; temperature max 69 C; lowest available memory 1944 MiB; pass timing: the full-size network 9dc98871eb7fc101 (one dispatch a frame) 5.77 ms a frame in the report period in the sample (pipeline compile 15.7 s on a worker). Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | branch mma-unroll (sheep’s BunnyHood commit 01f9b690d, a cherry-pick of their e95c60a, cherry-picked onto Longshot main): its gpu/ change (head f1365aaf; emit_spirv_warp.cpp, mma_layout.h, mma_software.h and vk_coopmat.cpp byte-identical on the rebased head, the vk_shader.cpp stamp hunk the same lines) applied on main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed) as a renderer-only pack via scripts/dev_packs.py (longshot_vk.so sha256 28bddba5f5a1b7c307f299ec170dc5ea8de75ad1b84d747920281c12998e950e, the game binary main’s); the same pack in every run of both arms |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2-2-neptune-72), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), stock automatic GPU clock | Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared (MemTotal 14.45 GiB) | native: 1280x800 output, render resolution 100% (options output_res=1280x800, resolution=1, upscaling=2), the game’s own upscaler, frame generation off; no longshot_env.txt; empty program and pipeline caches in a fresh private game folder (the translator changed), programs translated as first drawn; switches: none besides RX_VK_PRECOMPILE=0 (the default: the unrolled form tried on RADV); RX_VK_PASS_TIMING=2 (profiled: absolute times inflated) |
the main outdoor area (child, teleported from the starting-area save), 75 s warm-up then 45 s standing; one session 2026-11-16 04:54-05:46 UTC, timed runs interleaved loop, unrolled, unrolled, loop, then pass timing loop, unrolled, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy and APU power from sysfs at 10 Hz |
run mu-pt-U-fN, started 05:17:17 UTC, arm unrolled (default): log once vk: matrix multiplies in software (no cooperative matrices) and vk: software matrix multiplies unrolled, f16 rounding by conversion, A through quad broadcasts: the loop's words for 8192 lanes (RX_VK_MMA_UNROLL=0: looped). Profiled: 15.90 fps, 1% low 12.84 fps; present interval p50 / p95 / p99 62.6 / 66.8 / 71.2 ms, 1 over 100 ms; GPU time 40.5 ms a frame; GPU busy 75%, 13% of samples at 200 MHz, the rest mostly 1040 MHz; APU power mean / max 7.91 / 10.57 W; temperature max 62 C; lowest available memory 1931 MiB; pass timing: the full-size network 9dc98871eb7fc101 (one dispatch a frame) 4.30 and 4.05 ms a frame in the report periods after its compile (pipeline compile 15.3 s on a worker). Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | branch mma-unroll (sheep’s BunnyHood commit 01f9b690d, a cherry-pick of their e95c60a, cherry-picked onto Longshot main): its gpu/ change (head f1365aaf; emit_spirv_warp.cpp, mma_layout.h, mma_software.h and vk_coopmat.cpp byte-identical on the rebased head, the vk_shader.cpp stamp hunk the same lines) applied on main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed) as a renderer-only pack via scripts/dev_packs.py (longshot_vk.so sha256 28bddba5f5a1b7c307f299ec170dc5ea8de75ad1b84d747920281c12998e950e, the game binary main’s); the same pack in every run of both arms |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2-2-neptune-72), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), stock automatic GPU clock | Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared (MemTotal 14.45 GiB) | native: 1280x800 output, render resolution 100% (options output_res=1280x800, resolution=1, upscaling=2), the game’s own upscaler, frame generation off; no longshot_env.txt; empty program and pipeline caches in a fresh private game folder (the translator changed), programs translated as first drawn; switches: RX_VK_MMA_UNROLL=0 (the loop) and RX_VK_PRECOMPILE=0; RX_VK_PASS_TIMING=2 (profiled: absolute times inflated) |
the starting forest (child, the save’s spot outside the player character’s house beside Saria, about 2,200 draws a frame), 75 s warm-up then 45 s standing; one session 2026-11-16 04:54-05:46 UTC, timed runs interleaved loop, unrolled, unrolled, loop, then pass timing loop, unrolled, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy and APU power from sysfs at 10 Hz |
run mu-pt-L-kN, started 05:38:01 UTC, arm loop: log once vk: matrix multiplies in software (no cooperative matrices) and vk: software matrix multiplies looped (RX_VK_MMA_UNROLL=0). Profiled: 9.56 fps, 1% low 6.18 fps; present interval p50 / p95 / p99 101.8 / 136.6 / 152.7 ms, 249 over 100 ms; GPU time 72.5 ms a frame; GPU busy 78%, 21% of samples at 200 MHz, the rest mostly 1040 MHz; APU power mean / max 14.97 / 18.42 W; temperature max 70 C; lowest available memory 2428 MiB; pass timing: the full-size network 9dc98871eb7fc101 (one dispatch a frame) 9.48 ms a frame in the report period in the sample (pipeline compile 12.2 s on a worker). Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | branch mma-unroll (sheep’s BunnyHood commit 01f9b690d, a cherry-pick of their e95c60a, cherry-picked onto Longshot main): its gpu/ change (head f1365aaf; emit_spirv_warp.cpp, mma_layout.h, mma_software.h and vk_coopmat.cpp byte-identical on the rebased head, the vk_shader.cpp stamp hunk the same lines) applied on main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed) as a renderer-only pack via scripts/dev_packs.py (longshot_vk.so sha256 28bddba5f5a1b7c307f299ec170dc5ea8de75ad1b84d747920281c12998e950e, the game binary main’s); the same pack in every run of both arms |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2-2-neptune-72), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), stock automatic GPU clock | Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared (MemTotal 14.45 GiB) | native: 1280x800 output, render resolution 100% (options output_res=1280x800, resolution=1, upscaling=2), the game’s own upscaler, frame generation off; no longshot_env.txt; empty program and pipeline caches in a fresh private game folder (the translator changed), programs translated as first drawn; switches: RX_VK_MMA_UNROLL=0 (the loop) and RX_VK_PRECOMPILE=0; RX_VK_PASS_TIMING=2 (profiled: absolute times inflated) |
the main outdoor area (child, teleported from the starting-area save), 75 s warm-up then 45 s standing; one session 2026-11-16 04:54-05:46 UTC, timed runs interleaved loop, unrolled, unrolled, loop, then pass timing loop, unrolled, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy and APU power from sysfs at 10 Hz |
run mu-pt-L-fN, started 05:12:51 UTC, arm loop: log once vk: matrix multiplies in software (no cooperative matrices) and vk: software matrix multiplies looped (RX_VK_MMA_UNROLL=0). Profiled: 14.52 fps, 1% low 12.65 fps; present interval p50 / p95 / p99 68.7 / 72.9 / 76.7 ms, 0 over 100 ms; GPU time 45.6 ms a frame; GPU busy 76%, 11% of samples at 200 MHz, the rest mostly 1040 MHz; APU power mean / max 8.07 / 13.10 W; temperature max 63 C; lowest available memory 1583 MiB; pass timing: the full-size network 9dc98871eb7fc101 (one dispatch a frame) 9.51 and 9.27 ms a frame in the report periods after its compile (pipeline compile 12.0 s on a worker). Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | branch mma-unroll (sheep’s BunnyHood commit 01f9b690d, a cherry-pick of their e95c60a, cherry-picked onto Longshot main): its gpu/ change (head f1365aaf; emit_spirv_warp.cpp, mma_layout.h, mma_software.h and vk_coopmat.cpp byte-identical on the rebased head, the vk_shader.cpp stamp hunk the same lines) applied on main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed) as a renderer-only pack via scripts/dev_packs.py (longshot_vk.so sha256 28bddba5f5a1b7c307f299ec170dc5ea8de75ad1b84d747920281c12998e950e, the game binary main’s); the same pack in every run of both arms |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2-2-neptune-72), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), stock automatic GPU clock | Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared (MemTotal 14.45 GiB) | native: 1280x800 output, render resolution 100% (options output_res=1280x800, resolution=1, upscaling=2), the game’s own upscaler, frame generation off; no longshot_env.txt; empty program and pipeline caches in a fresh private game folder (the translator changed), programs translated as first drawn; switches: RX_VK_MMA_UNROLL=0 (the loop) and RX_VK_PRECOMPILE=0 |
the starting forest (child, the save’s spot outside the player character’s house beside Saria, about 2,200 draws a frame), 75 s warm-up then 45 s standing; one session 2026-11-16 04:54-05:46 UTC, timed runs interleaved loop, unrolled, unrolled, loop, then pass timing loop, unrolled, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy and APU power from sysfs at 10 Hz |
run mu-L-kN-2, started 05:33:57 UTC, arm loop: log once vk: matrix multiplies in software (no cooperative matrices) and vk: software matrix multiplies looped (RX_VK_MMA_UNROLL=0). 9.20 fps, 1% low 6.24 fps; present interval p50 / p95 / p99 101.7 / 148.7 / 153.8 ms, 240 over 100 ms; GPU time 72.6 ms a frame; GPU busy 86%, 20% of samples at 200 MHz, the rest mostly 1040 MHz; APU power mean / max 15.01 / 18.37 W; temperature max 69 C; lowest available memory 2002 MiB. Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | branch mma-unroll (sheep’s BunnyHood commit 01f9b690d, a cherry-pick of their e95c60a, cherry-picked onto Longshot main): its gpu/ change (head f1365aaf; emit_spirv_warp.cpp, mma_layout.h, mma_software.h and vk_coopmat.cpp byte-identical on the rebased head, the vk_shader.cpp stamp hunk the same lines) applied on main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed) as a renderer-only pack via scripts/dev_packs.py (longshot_vk.so sha256 28bddba5f5a1b7c307f299ec170dc5ea8de75ad1b84d747920281c12998e950e, the game binary main’s); the same pack in every run of both arms |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2-2-neptune-72), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), stock automatic GPU clock | Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared (MemTotal 14.45 GiB) | native: 1280x800 output, render resolution 100% (options output_res=1280x800, resolution=1, upscaling=2), the game’s own upscaler, frame generation off; no longshot_env.txt; empty program and pipeline caches in a fresh private game folder (the translator changed), programs translated as first drawn; switches: RX_VK_MMA_UNROLL=0 (the loop) and RX_VK_PRECOMPILE=0 |
the starting forest (child, the save’s spot outside the player character’s house beside Saria, about 2,200 draws a frame), 75 s warm-up then 45 s standing; one session 2026-11-16 04:54-05:46 UTC, timed runs interleaved loop, unrolled, unrolled, loop, then pass timing loop, unrolled, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy and APU power from sysfs at 10 Hz |
run mu-L-kN-1, started 05:21:44 UTC, arm loop: log once vk: matrix multiplies in software (no cooperative matrices) and vk: software matrix multiplies looped (RX_VK_MMA_UNROLL=0). 9.12 fps, 1% low 5.89 fps; present interval p50 / p95 / p99 103.4 / 149.1 / 159.5 ms, 247 over 100 ms; GPU time 79.1 ms a frame; GPU busy 83%, 22% of samples at 200 MHz, the rest mostly 1040 MHz; APU power mean / max 14.99 / 18.96 W; temperature max 69 C; lowest available memory 2016 MiB. Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | branch mma-unroll (sheep’s BunnyHood commit 01f9b690d, a cherry-pick of their e95c60a, cherry-picked onto Longshot main): its gpu/ change (head f1365aaf; emit_spirv_warp.cpp, mma_layout.h, mma_software.h and vk_coopmat.cpp byte-identical on the rebased head, the vk_shader.cpp stamp hunk the same lines) applied on main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed) as a renderer-only pack via scripts/dev_packs.py (longshot_vk.so sha256 28bddba5f5a1b7c307f299ec170dc5ea8de75ad1b84d747920281c12998e950e, the game binary main’s); the same pack in every run of both arms |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2-2-neptune-72), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), stock automatic GPU clock | Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared (MemTotal 14.45 GiB) | native: 1280x800 output, render resolution 100% (options output_res=1280x800, resolution=1, upscaling=2), the game’s own upscaler, frame generation off; no longshot_env.txt; empty program and pipeline caches in a fresh private game folder (the translator changed), programs translated as first drawn; switches: RX_VK_MMA_UNROLL=0 (the loop) and RX_VK_PRECOMPILE=0 |
the main outdoor area (child, teleported from the starting-area save), 75 s warm-up then 45 s standing; one session 2026-11-16 04:54-05:46 UTC, timed runs interleaved loop, unrolled, unrolled, loop, then pass timing loop, unrolled, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy and APU power from sysfs at 10 Hz |
run mu-L-fN-2, started 05:08:23 UTC, arm loop: log once vk: matrix multiplies in software (no cooperative matrices) and vk: software matrix multiplies looped (RX_VK_MMA_UNROLL=0). 14.76 fps, 1% low 12.81 fps; present interval p50 / p95 / p99 67.3 / 73.6 / 76.5 ms, 0 over 100 ms; GPU time 45.5 ms a frame; GPU busy 75%, 12% of samples at 200 MHz, the rest mostly 1040 MHz; APU power mean / max 8.11 / 10.27 W; temperature max 63 C; lowest available memory 1999 MiB. Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | branch mma-unroll (sheep’s BunnyHood commit 01f9b690d, a cherry-pick of their e95c60a, cherry-picked onto Longshot main): its gpu/ change (head f1365aaf; emit_spirv_warp.cpp, mma_layout.h, mma_software.h and vk_coopmat.cpp byte-identical on the rebased head, the vk_shader.cpp stamp hunk the same lines) applied on main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed) as a renderer-only pack via scripts/dev_packs.py (longshot_vk.so sha256 28bddba5f5a1b7c307f299ec170dc5ea8de75ad1b84d747920281c12998e950e, the game binary main’s); the same pack in every run of both arms |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2-2-neptune-72), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), stock automatic GPU clock | Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared (MemTotal 14.45 GiB) | native: 1280x800 output, render resolution 100% (options output_res=1280x800, resolution=1, upscaling=2), the game’s own upscaler, frame generation off; no longshot_env.txt; empty program and pipeline caches in a fresh private game folder (the translator changed), programs translated as first drawn; switches: RX_VK_MMA_UNROLL=0 (the loop) and RX_VK_PRECOMPILE=0 |
the main outdoor area (child, teleported from the starting-area save), 75 s warm-up then 45 s standing; one session 2026-11-16 04:54-05:46 UTC, timed runs interleaved loop, unrolled, unrolled, loop, then pass timing loop, unrolled, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy and APU power from sysfs at 10 Hz |
run mu-L-fN-1, started 04:54:58 UTC, arm loop: log once vk: matrix multiplies in software (no cooperative matrices) and vk: software matrix multiplies looped (RX_VK_MMA_UNROLL=0). 14.89 fps, 1% low 12.98 fps; present interval p50 / p95 / p99 67.0 / 71.2 / 75.1 ms, 0 over 100 ms; GPU time 45.4 ms a frame; GPU busy 74%, 15% of samples at 200 MHz, the rest mostly 1040 MHz; APU power mean / max 8.02 / 9.64 W; temperature max 63 C; lowest available memory 1347 MiB. Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 | main f7922126 and topic 4155b515 | Windows 10 Pro 19045 | RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT; 64 GB | First comparison attempt; copied settings; windowed mode | rejected-window-comparison; stationary the village | Harness rejected differing observed window sizes before accepting the topic comparison. Partial FPS data excluded from conclusions. Corrected fixed-size runs are separately recorded. |
| 2026-11-16 | feat/microphone-test-and-gain 4240dd3ac2afd768b2d11f94896fad6a9fea230a plus uncommitted source-archive SHA-256 dc13c56616421df3c7e3dcb3f673d87b23f8a488b0186029151264a33408b3e9; matching ABI-25 host and renderer | Windows 10 Pro 19045; desktop | RTX 3080; WMI driver 32.0.16.2124 | Ryzen 9 5900XT; 64 GB | Existing per-PC microphone, gain, audio/layout and control settings; explicit reusable development user folder; normal cache checks | windows-shared-user-01; manual launch and owner inspection of saves/settings; exact game scene not recorded | Owner confirmed expected saves/progress and settings correct. Log confirmed the intended options file and save mount; 47867 program keys reused; no tagged error/fatal, VK_ERROR or not-translated matches in the observed log. Initial loading verified; no subsequent changed-setting restart check or performance comparison in this run. |
| 2026-11-16 | feat/microphone-test-and-gain 4240dd3ac2afd768b2d11f94896fad6a9fea230a plus uncommitted source-archive SHA-256 dc13c56616421df3c7e3dcb3f673d87b23f8a488b0186029151264a33408b3e9 | Windows; OS build not recorded; no window | Not used; game-free fixtures | Not recorded for this fixture run | Python 3.14; llvm-mingw Clang 23.1.5; fake audio/UI; tests/run_all.py, 4 jobs; MAGICBEAN_TOOLING set for retries/controls, unset in full initial run | Selected game-free baseline/setup control; no gameplay | windows-retry: 2 passed, 23 failed, 0 skipped in 22.18 s. Every initial suite failure rerun with the existing shared Tooling path set; two setup failures cleared. Remaining failures match pinned main. Source hashes unchanged; test durations are diagnostic, not game-performance measurements. |
| 2026-11-16 | f2709248 plus six source/helper edits, subsequently committed as 4cdf7354 | Windows 10 Pro 19045 | RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT; 64 GB | Reusable independent dev saves/settings; changed microphone gain/device; normal close/reopen | owner-reported-persistence; manual verification before controlled comparisons | Owner confirmed changed gain/device persistence across restart on this PC, then closed the game. No every-slot or cross-PC synchronization claim. |
| 2026-11-16 | feat/microphone-test-and-gain f2709248367dc697a8044cf530579b1ece05a57c plus six uncommitted source/helper edits; runtime source eb356c3c; matching ABI-25 host/renderer | Windows 10 Pro 19045; desktop game launch | RTX 3080; WMI driver 32.0.16.2124 | Ryzen 9 5900XT; 64 GB | Existing reusable dev saves/settings; normal-optimization build; normal cache checks; actual display 2560x1440 at 120 Hz; no controlled performance settings | windows-rebased-play-01; manual Audio controls, save/progress and settings check; exact gameplay scene not recorded | Owner confirmed all requested checks pass: expected saves/settings, microphone selection, Start/Stop, meters on Test/Gain, gain adjustment, playback/Mono and stopping the test on Audio/focus exit. Logs confirm the intended options/save paths, accepted Auto settings and 47867 retained program keys; no tagged errors in observed snapshot. No deliberate new-setting restart, exhaustive device edge cases, screenshots or performance comparison claimed. |
| 2026-11-16 | main f792212691eb3d49a37ff90b5690f0da20077a41 | Windows 10 Pro 19045; no window or hardware capture | RTX 3080; WMI driver 32.0.16.2124; GPU not used by these fixtures | Ryzen 9 5900XT; 64 GB | Python build venv 3.14.5 with zstandard; llvm-mingw Clang 23.1.5; existing headers and Tooling; 4 fixture jobs | windows-main-controls; All failing topic checks selected against pinned main; no full main suite | 0 passed, 25 failed, 0 skipped in 113.48 s including runner overhead. All 25 topic failures reproduced; 24 normalized signatures match. Control-helper has shared relaunch/pipe failures plus extra main cleanup errors. Source hashes unchanged; no game build/launch or performance claim. |
| 2026-11-16 | feat/microphone-test-and-gain f2709248367dc697a8044cf530579b1ece05a57c plus six uncommitted source/helper edits; source-only transfer archive SHA-256 e15c1195cf9e2e7bc8e78eb8d811e96d28d0980d864877132ef4756ece844b14 | Windows 10 Pro 19045; no window or hardware capture | RTX 3080; WMI driver 32.0.16.2124; GPU not used by these fixtures | Ryzen 9 5900XT; 64 GB | Python build venv 3.14.5 with zstandard; llvm-mingw Clang 23.1.5; existing headers and Tooling; 4 fixture jobs | windows-full; Full maintained game-free suite, 667 checks; no game dump | 600 passed, 25 failed, 42 skipped in 323.86 s including runner overhead. All 25 failure names reproduce on main; 24 diagnostic signatures match, control-helper primary failures are shared with differing cleanup diagnostics. Source hashes unchanged; no game build/launch or performance claim. |
| 2026-11-16 | feat/microphone-test-and-gain f2709248367dc697a8044cf530579b1ece05a57c plus six uncommitted source/helper edits; source-only transfer archive SHA-256 e15c1195cf9e2e7bc8e78eb8d811e96d28d0980d864877132ef4756ece844b14 | Windows 10 Pro 19045; no window or hardware capture | RTX 3080; WMI driver 32.0.16.2124; GPU not used by these fixtures | Ryzen 9 5900XT; 64 GB | Python build venv 3.14.5 with zstandard; llvm-mingw Clang 23.1.5; existing headers and Tooling; 4 fixture jobs | windows-focus-final; Expanded mic/audio, frame-cap, fps, build-recovery, crash, release and path fixtures | 33 passed, 3 failed, 1 skipped in 25.25 s including runner overhead. Core mic/audio fixtures pass. Expanded failures also reproduce on pinned main; exact reasons remain in authored test logs. Source hashes unchanged; no game build/launch or performance claim. |
| 2026-11-16 | feat/microphone-test-and-gain f2709248367dc697a8044cf530579b1ece05a57c plus six uncommitted source/helper edits; source-only transfer archive SHA-256 e15c1195cf9e2e7bc8e78eb8d811e96d28d0980d864877132ef4756ece844b14 | Windows 10 Pro 19045; no window or hardware capture | RTX 3080; WMI driver 32.0.16.2124; GPU not used by these fixtures | Ryzen 9 5900XT; 64 GB | Python build venv 3.14.5 with zstandard; llvm-mingw Clang 23.1.5; existing headers and Tooling; 4 fixture jobs | windows-focus; Expanded mic/audio, frame-cap, fps, build-recovery, crash, release and path fixtures | 32 passed, 4 failed, 0 skipped in 29.23 s including runner overhead. Superseded invocation: explicitly named POSIX-only save fixture and omitted the fps Python driver; corrected focus is recorded separately. Source hashes unchanged; no game build/launch or performance claim. |
| 2026-11-16 | feat/microphone-test-and-gain f2709248367dc697a8044cf530579b1ece05a57c plus six uncommitted source/helper edits; runtime source eb356c3c; matching ABI-25 host/renderer | Windows 10 Pro 19045; no gameplay window | RTX 3080; WMI driver 32.0.16.2124 | Ryzen 9 5900XT; 64 GB | Existing build Python 3.14.5 and llvm-mingw Clang 23.1.5; normal optimization; retained work; 12 compiler jobs; Most headless preparation with normal cache checks | windows-rebased-build-01; incremental update, static PE/dependency inspection, then shader preparation only | Update exited 0 in 23.60 s; native compile 7 s, zero failed; 7 new and 505 cached file reports. Static inspection passed. Preparation exited 0 in 11.50 s, retained 47867 program keys, prepared 33269 noted graphics and 9927 compute pipelines with none missing. Saves/settings, launchers and protected retained binaries match pre-update hashes. No gameplay or fps claim. |
| 2026-11-16 | Mic branch c314e05e on main 3bb97bef; final affected rerun | Windows 10 Pro 19045; game-free | RTX 3080; driver 32.0.16.2124; not used by fixtures | Ryzen 9 5900XT; 64 GB | Installed LLVM 23.1.5 and isolated Python; nine affected checks; four jobs | Rebase-ready affected mic, gain, help, translation, menu keys, ABI, framegen-provider and upscaler checks, including the final small-card DLSS cases | Eight pass; test_bug_report_defaults fails its existing auto-profile shadow_res assertion, reproduced on current main. Rebase conflict was translation counts only; capture, host wiring, ABI and paragraph code unchanged. Historical measurements keep original builds; final sub-7-GB DLSS default does not apply to the measured 10/24-GB GPUs. No game build or launch. |
| 2026-11-16 | bc0997d444f951445653a7d7aa49f78471a18a7e | Windows 10 Pro 19045 | RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT; 64 GB | 2560x1440; Auto settled render at 67%; same Test microphone help view; gain 100%; frame generation off; renderer cap 60; normal optimization | overlay-testing-02; stationary the village save, same process and selected row; interleaved capture states; 25 s sample window after settling | Native FPS median 50.85 from 10 intervals; cumulative capture-open log entries 2. Clock anchored to the first observed microphone-open event; tested/stopped/tested/stopped. Samples vary; no performance gain or zero-cost claim. |
| 2026-11-16 | bc0997d444f951445653a7d7aa49f78471a18a7e | Windows 10 Pro 19045 | RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT; 64 GB | 2560x1440; Auto settled render at 67%; same Test microphone help view; gain 100%; frame generation off; renderer cap 60; normal optimization | overlay-testing-01; stationary the village save, same process and selected row; interleaved capture states; 25 s sample window after settling | Native FPS median 50.30 from 10 intervals; cumulative capture-open log entries 1. Clock anchored to the first observed microphone-open event; tested/stopped/tested/stopped. Samples vary; no performance gain or zero-cost claim. |
| 2026-11-16 | bc0997d444f951445653a7d7aa49f78471a18a7e | Windows 10 Pro 19045 | RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT; 64 GB | 2560x1440; Auto settled render at 67%; same Test microphone help view; gain 100%; frame generation off; renderer cap 60; normal optimization | overlay-stopped-02; stationary the village save, same process and selected row; interleaved capture states; 25 s sample window after settling | Native FPS median 52.35 from 10 intervals; cumulative capture-open log entries 2. Clock anchored to the first observed microphone-open event; tested/stopped/tested/stopped. Samples vary; no performance gain or zero-cost claim. |
| 2026-11-16 | bc0997d444f951445653a7d7aa49f78471a18a7e | Windows 10 Pro 19045 | RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT; 64 GB | 2560x1440; Auto settled render at 67%; same Test microphone help view; gain 100%; frame generation off; renderer cap 60; normal optimization | overlay-stopped-01; stationary the village save, same process and selected row; interleaved capture states; 25 s sample window after settling | Native FPS median 49.30 from 9 intervals; cumulative capture-open log entries 1. Clock anchored to the first observed microphone-open event; tested/stopped/tested/stopped. Samples vary; no performance gain or zero-cost claim. |
| 2026-11-16 | bc0997d444f951445653a7d7aa49f78471a18a7e | Windows 10 Pro 19045 | RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT; 64 GB | First native-menu overlay timing attempt; copied data; no profiling | rejected-overlay-clock; first-present clock anchor | Capture-open counters showed phase labels were not aligned to the native menu clock. Entire comparison excluded; no stopped/active verdict from these samples. The replacement uses a live first-capture-open anchor. |
| 2026-11-16 | main bbe421f508a3aa6c32014265067a5031ef333b32 | Windows; OS build not recorded; no window | Not used; game-free fixtures | Not recorded for this fixture run | Python 3.14; llvm-mingw Clang 23.1.5; fake audio/UI; tests/run_all.py, 4 jobs; MAGICBEAN_TOOLING set for retries/controls, unset in full initial run | Selected game-free baseline/setup control; no gameplay | windows-main: 2 passed, 23 failed, 0 skipped in 22.45 s. Selected main control for every initial topic failure; two Tooling-dependent checks passed after setup correction. All remaining failure names and diagnostic signatures match the topic. Source hashes unchanged; test durations are diagnostic, not game-performance measurements. |
| 2026-11-16 | main bbe421f508a3aa6c32014265067a5031ef333b32 | Windows; OS build not recorded; no window | Not used; game-free fixtures | Not recorded for this fixture run | Python 3.14; llvm-mingw Clang 23.1.5; fake audio/UI; tests/run_all.py, 4 jobs; existing tools only | Selected game-free baseline/setup control; no gameplay | windows-main-initial: 0 passed, 12 failed, 0 skipped in 5.22 s. Partial control before Tooling-path correction; all 12 selected names failed. Superseded by windows-main. Source hashes unchanged; test durations are diagnostic, not game-performance measurements. |
| 2026-11-16 | feat/microphone-test-and-gain 4240dd3ac2afd768b2d11f94896fad6a9fea230a plus uncommitted source-archive SHA-256 dc13c56616421df3c7e3dcb3f673d87b23f8a488b0186029151264a33408b3e9 | Windows; OS build not recorded; no window | Not used; game-free fixtures | Not recorded for this fixture run | Python 3.14; llvm-mingw Clang 23.1.5; fake audio/UI; tests/run_all.py, 4 jobs; MAGICBEAN_TOOLING set for retries/controls, unset in full initial run | Complete maintained game-free suite, 666 checks including skips; no gameplay | windows-full: 592 passed, 25 failed, 49 skipped in 291.49 s. Complete suite; two setup failures from an unset shared Tooling path cleared in the recorded retry. Remaining failures reproduce on pinned main. Source hashes unchanged; test durations are diagnostic, not game-performance measurements. |
| 2026-11-16 | feat/microphone-test-and-gain 4240dd3ac2afd768b2d11f94896fad6a9fea230a plus uncommitted source-archive SHA-256 dc13c56616421df3c7e3dcb3f673d87b23f8a488b0186029151264a33408b3e9 | Windows; OS build not recorded; no window | Not used; game-free fixtures | Not recorded for this fixture run | Python 3.14; llvm-mingw Clang 23.1.5; fake audio/UI; 4 concurrent fixtures; existing tools only | Focused 30-script mic/audio selection; no window, GPU or hardware capture | windows-focus: 29 passed, 0 failed, 1 skipped in 9.38 s. All portable mic/audio fixtures passed; the project’s POSIX-only atomic-save skip was respected. Source hashes unchanged; test durations are diagnostic, not game-performance measurements. |
| 2026-11-16 | feat/microphone-test-and-gain 4240dd3ac2afd768b2d11f94896fad6a9fea230a plus uncommitted source-archive SHA-256 dc13c56616421df3c7e3dcb3f673d87b23f8a488b0186029151264a33408b3e9 | Windows; OS build not recorded; no window | Not used; game-free fixtures | Not recorded for this fixture run | Python 3.14; llvm-mingw Clang 23.1.5; fake audio/UI; 4 concurrent fixtures; existing tools only | Focused 30-script mic/audio selection; no window, GPU or hardware capture | windows-focus-initial: 29 passed, 1 failed, 0 skipped in 14.15 s. Runner incorrectly invoked the declared POSIX-only save fixture on Windows; os.geteuid failed before compilation. Superseded by windows-focus; no production failure. Source hashes unchanged; test durations are diagnostic, not game-performance measurements. |
| 2026-11-16 | bc0997d444f951445653a7d7aa49f78471a18a7e plus reviewed guide documentation | Windows 10 Pro 19045 | RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT; 64 GB | Installed Clang 23 and isolated Python; four fixture jobs; normal OS skips; source hashes verified unchanged | final-full-suite; maintained tests/run_all.py with JUnit; game-free only | 600 passed, 25 failed, 42 skipped, 667 checks in 322.91 s. All 25 failed names match the prior topic and pinned-main controls; no new failing names. No game build or launch. |
| 2026-11-16 | bc0997d444f951445653a7d7aa49f78471a18a7e | Windows 10 Pro 19045 | RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT; 64 GB | 2560x1440; Auto settled render at 67%; normal optimization; gain 100%; test stopped; frame generation off; renderer cap 60; V-Sync Auto/mailbox; host HOS_FPS_MAX=1000 is a separate ceiling | default-topic-02; stationary KakarikoVillage_Child from identical copied save; 45 s warm-up per side, 30 s settling and 60 s measurement; main/topic/main/topic | Native FPS median 50.55 from 24 log intervals; observed sample range 48.5-52.2 FPS. Matching scene, position, camera and window verified. This capped test is not proof of zero CPU cost. |
| 2026-11-16 | bc0997d444f951445653a7d7aa49f78471a18a7e | Windows 10 Pro 19045 | RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT; 64 GB | 2560x1440; Auto settled render at 67%; normal optimization; gain 100%; test stopped; frame generation off; renderer cap 60; V-Sync Auto/mailbox; host HOS_FPS_MAX=1000 is a separate ceiling | default-topic-01; stationary KakarikoVillage_Child from identical copied save; 45 s warm-up per side, 30 s settling and 60 s measurement; main/topic/main/topic | Native FPS median 50.40 from 24 log intervals; observed sample range 42.0-51.9 FPS. Matching scene, position, camera and window verified. This capped test is not proof of zero CPU cost. |
| 2026-11-16 | f792212691eb3d49a37ff90b5690f0da20077a41 | Windows 10 Pro 19045 | RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT; 64 GB | 2560x1440; Auto settled render at 67%; normal optimization; gain 100%; test stopped; frame generation off; renderer cap 60; V-Sync Auto/mailbox; host HOS_FPS_MAX=1000 is a separate ceiling | default-main-02; stationary KakarikoVillage_Child from identical copied save; 45 s warm-up per side, 30 s settling and 60 s measurement; main/topic/main/topic | Native FPS median 50.70 from 24 log intervals; observed sample range 40.7-52.1 FPS. Matching scene, position, camera and window verified. This capped test is not proof of zero CPU cost. |
| 2026-11-16 | f792212691eb3d49a37ff90b5690f0da20077a41 | Windows 10 Pro 19045 | RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT; 64 GB | 2560x1440; Auto settled render at 67%; normal optimization; gain 100%; test stopped; frame generation off; renderer cap 60; V-Sync Auto/mailbox; host HOS_FPS_MAX=1000 is a separate ceiling | default-main-01; stationary KakarikoVillage_Child from identical copied save; 45 s warm-up per side, 30 s settling and 60 s measurement; main/topic/main/topic | Native FPS median 50.40 from 24 log intervals; observed sample range 46.3-51.4 FPS. Matching scene, position, camera and window verified. This capped test is not proof of zero CPU cost. |
| 2026-11-16 | main f7922126 and topic 4155b515 | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; GNOME desktop | RTX 5090 Laptop GPU; NVIDIA 595.99.02 | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | First comparison attempt; copied settings; windowed mode | rejected-window-comparison; stationary the village | Harness rejected differing observed window sizes before accepting the topic comparison. Partial FPS data excluded from conclusions. Corrected fixed-size runs are separately recorded. |
| 2026-11-16 | f792212691eb3d49a37ff90b5690f0da20077a41 before; bc0997d444f951445653a7d7aa49f78471a18a7e after | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; GNOME desktop | RTX 5090 Laptop GPU; NVIDIA 595.99.02 | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | Game-free SDL software renderer; maintained menu/ImGui and bundled Libertinus; English 1280x720 plus German 960x720; Medium size | ui-only-guide-and-mic-help; no game/GPU/microphone/save opened | Before/after Audio and synthetic active/stopped Test/Gain views captured and visually reviewed. German final-help clipping fixed and recaptured with full paragraphs/bullets/clipping note. Canonical Graphics image/provenance updated; offline source checks pass. Synthetic 24%/48% meter values prove layout only. |
| 2026-11-16 | feat/microphone-test-and-gain 4240dd3ac2afd768b2d11f94896fad6a9fea230a plus uncommitted source-archive SHA-256 dc13c56616421df3c7e3dcb3f673d87b23f8a488b0186029151264a33408b3e9; matching ABI-25 host and renderer | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; existing graphical desktop | RTX 5090 Laptop GPU; NVIDIA 595.99.02 | Core Ultra 9 275HX; 64 GB installed class, about 61.2 GiB kernel-visible | Existing per-PC microphone, gain, audio/layout and control settings; explicit reusable development user folder; normal cache checks | linux-shared-user-01; manual owner inspection after desktop-session launch; exact game scene not recorded | Owner confirmed expected saves/progress and settings correct. Log confirmed the intended options file and save mount; 47867 program keys reused; no tagged error/fatal, VK_ERROR or not-translated matches in the observed log. Game subsequently closed normally, exit 0. Initial loading verified; no subsequent changed-setting restart check or performance comparison in this run. |
| 2026-11-16 | feat/microphone-test-and-gain 4240dd3ac2afd768b2d11f94896fad6a9fea230a plus uncommitted source-archive SHA-256 dc13c56616421df3c7e3dcb3f673d87b23f8a488b0186029151264a33408b3e9 | Native Ubuntu laptop; distro/kernel version not recorded for this run; no window | Not used; game-free fixtures | Not recorded for this fixture run | Python 3.14 (isolated environment); LLVM 23.1.6; fake audio/UI; tests/run_all.py, 4 jobs; MAGICBEAN_TOOLING set for retries/controls, unset in full initial run | Selected game-free baseline/setup control; no gameplay | linux-retry: 2 passed, 18 failed, 0 skipped in 27.45 s. Every initial suite failure rerun with the existing shared Tooling path set; two setup failures cleared. Remaining failures match pinned main. Source hashes unchanged; test durations are diagnostic, not game-performance measurements. |
| 2026-11-16 | f2709248 plus six source/helper edits, subsequently committed as 4cdf7354 | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; GNOME desktop | RTX 5090 Laptop GPU; NVIDIA 595.99.02 | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | Reusable independent dev saves/settings; changed microphone gain/device; normal close/reopen | owner-reported-persistence; manual verification before controlled comparisons | Owner confirmed changed gain/device persistence across restart on this PC, then closed the game. No every-slot or cross-PC synchronization claim. |
| 2026-11-16 | feat/microphone-test-and-gain f2709248367dc697a8044cf530579b1ece05a57c plus six uncommitted source/helper edits; runtime source eb356c3c; matching ABI-25 host/renderer | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; existing graphical desktop session | RTX 5090 Laptop GPU; NVIDIA 595.99.02 | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | Existing reusable dev saves/settings; normal-optimization build; normal cache checks; actual display 1920x1200 at 240 Hz; no controlled performance settings | linux-rebased-play-01; manual Audio controls, save/progress and settings check; exact gameplay scene not recorded | Owner confirmed all requested checks pass: expected saves/settings, microphone selection, Start/Stop, meters on Test/Gain, gain adjustment, playback/Mono and stopping the test on Audio/focus exit. Logs confirm the intended options/save paths, accepted Auto settings and 47867 retained program keys; no tagged errors in observed snapshot. No deliberate new-setting restart, exhaustive device edge cases, screenshots or performance comparison claimed. |
| 2026-11-16 | main f792212691eb3d49a37ff90b5690f0da20077a41 | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; no window or hardware capture | RTX 5090 Laptop GPU; NVIDIA 595.99.02; GPU not used by these fixtures | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | Isolated Python 3.14.4; LLVM 23.1.6; existing headers and Tooling; 4 fixture jobs | linux-main-controls; All failing topic checks selected against pinned main; no full main suite | 0 passed, 18 failed, 0 skipped in 13.96 s including runner overhead. All 18 topic failures reproduced with matching normalized diagnostic signatures. Source hashes unchanged; no game build/launch or performance claim. |
| 2026-11-16 | feat/microphone-test-and-gain f2709248367dc697a8044cf530579b1ece05a57c plus six uncommitted source/helper edits; source-only transfer archive SHA-256 e15c1195cf9e2e7bc8e78eb8d811e96d28d0980d864877132ef4756ece844b14 | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; no window or hardware capture | RTX 5090 Laptop GPU; NVIDIA 595.99.02; GPU not used by these fixtures | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | Isolated Python 3.14.4; LLVM 23.1.6; existing headers and Tooling; 4 fixture jobs | linux-full; Full maintained game-free suite, 667 checks; no game dump | 632 passed, 18 failed, 17 skipped in 306.99 s including runner overhead. All 18 failures reproduce on main with matching normalized diagnostic signatures. Source hashes unchanged; no game build/launch or performance claim. |
| 2026-11-16 | feat/microphone-test-and-gain f2709248367dc697a8044cf530579b1ece05a57c plus six uncommitted source/helper edits; source-only transfer archive SHA-256 e15c1195cf9e2e7bc8e78eb8d811e96d28d0980d864877132ef4756ece844b14 | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; no window or hardware capture | RTX 5090 Laptop GPU; NVIDIA 595.99.02; GPU not used by these fixtures | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | Isolated Python 3.14.4; LLVM 23.1.6; existing headers and Tooling; 4 fixture jobs | linux-focus; Expanded mic/audio, frame-cap, fps, build-recovery, crash, release and path fixtures | 36 passed, 1 failed, 0 skipped in 16.54 s including runner overhead. Core mic/audio fixtures pass. Expanded failures also reproduce on pinned main; exact reasons remain in authored test logs. Source hashes unchanged; no game build/launch or performance claim. |
| 2026-11-16 | feat/microphone-test-and-gain f2709248367dc697a8044cf530579b1ece05a57c plus six uncommitted source/helper edits; runtime source eb356c3c; matching ABI-25 host/renderer | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; existing desktop graphics context; no gameplay window | RTX 5090 Laptop GPU; NVIDIA 595.99.02 | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | Isolated Python 3.14.4 and LLVM 23.1.6; normal optimization; retained work; 10 compiler jobs; Most headless preparation with normal cache checks | linux-rebased-build-01; incremental update, static ELF/dependency/version inspection, then desktop-user-manager shader preparation only | Update exited 0 in 20.54 s; native compile 17 s, zero failed; 39 new and 475 cached file reports. Static inspection passed. Preparation exited 0 in 6.48 s, retained 47867 program keys, prepared 33185 noted graphics and 9927 compute pipelines with none missing. Preparation changed only Auto-tuning markers; verified pre-preparation settings restored. Final saves/settings, launchers and protected retained binaries match original hashes. No gameplay or fps claim. |
| 2026-11-16 | Mic branch c314e05e on main 3bb97bef; final affected rerun | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; game-free | RTX 5090 Laptop GPU; NVIDIA 595.99.02; not used by fixtures | Core Ultra 9 275HX; about 61.2 GiB RAM | Installed LLVM 23.1.6 and isolated Python; nine affected checks; four jobs | Rebase-ready mic, gain, help, translation, menu keys, ABI, FG and upscaler checks, including final small-card DLSS cases; UI-only guide source review | Eight pass; existing auto-profile shadow_res assertion reproduces on current main. The inspected 5307877d UI-only Graphics capture is retained for the final below-view DLSS-model delta; original provenance preserved. Mic patch unchanged except combined counts. Historical measurements remain scoped; sub-7-GB default does not apply to measured GPUs. No full suite, game build or gameplay launch. |
| 2026-11-16 | bc0997d444f951445653a7d7aa49f78471a18a7e | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; GNOME desktop | RTX 5090 Laptop GPU; NVIDIA 595.99.02 | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | 2560x1600 reported window/swapchain after startup; Auto retains saved display choices; same Test microphone help view; gain 100%; frame generation off; renderer cap 60; normal optimization | overlay-testing-02; stationary the village save, same process and selected row; interleaved capture states; 25 s sample window after settling | Native FPS median 59.40 from 12 intervals; cumulative capture-open log entries 2. Stopped/tested/stopped/tested; phase-end capture-open counts 0,1,1,2 bound the native menu clock alignment. Samples vary; no performance gain or zero-cost claim. |
| 2026-11-16 | bc0997d444f951445653a7d7aa49f78471a18a7e | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; GNOME desktop | RTX 5090 Laptop GPU; NVIDIA 595.99.02 | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | 2560x1600 reported window/swapchain after startup; Auto retains saved display choices; same Test microphone help view; gain 100%; frame generation off; renderer cap 60; normal optimization | overlay-testing-01; stationary the village save, same process and selected row; interleaved capture states; 25 s sample window after settling | Native FPS median 60.10 from 12 intervals; cumulative capture-open log entries 1. Stopped/tested/stopped/tested; phase-end capture-open counts 0,1,1,2 bound the native menu clock alignment. Samples vary; no performance gain or zero-cost claim. |
| 2026-11-16 | bc0997d444f951445653a7d7aa49f78471a18a7e | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; GNOME desktop | RTX 5090 Laptop GPU; NVIDIA 595.99.02 | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | 2560x1600 reported window/swapchain after startup; Auto retains saved display choices; same Test microphone help view; gain 100%; frame generation off; renderer cap 60; normal optimization | overlay-stopped-02; stationary the village save, same process and selected row; interleaved capture states; 25 s sample window after settling | Native FPS median 56.40 from 11 intervals; cumulative capture-open log entries 1. Stopped/tested/stopped/tested; phase-end capture-open counts 0,1,1,2 bound the native menu clock alignment. Samples vary; no performance gain or zero-cost claim. |
| 2026-11-16 | bc0997d444f951445653a7d7aa49f78471a18a7e | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; GNOME desktop | RTX 5090 Laptop GPU; NVIDIA 595.99.02 | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | 2560x1600 reported window/swapchain after startup; Auto retains saved display choices; same Test microphone help view; gain 100%; frame generation off; renderer cap 60; normal optimization | overlay-stopped-01; stationary the village save, same process and selected row; interleaved capture states; 25 s sample window after settling | Native FPS median 59.95 from 12 intervals; cumulative capture-open log entries 0. Stopped/tested/stopped/tested; phase-end capture-open counts 0,1,1,2 bound the native menu clock alignment. Samples vary; no performance gain or zero-cost claim. |
| 2026-11-16 | main bbe421f508a3aa6c32014265067a5031ef333b32 | Native Ubuntu laptop; distro/kernel version not recorded for this run; no window | Not used; game-free fixtures | Not recorded for this fixture run | Python 3.14 (isolated environment); LLVM 23.1.6; fake audio/UI; tests/run_all.py, 4 jobs; MAGICBEAN_TOOLING set for retries/controls, unset in full initial run | Selected game-free baseline/setup control; no gameplay | linux-main: 2 passed, 18 failed, 0 skipped in 28.47 s. Selected main control for every initial topic failure; two Tooling-dependent checks passed after setup correction. All remaining failure names and diagnostic signatures match the topic. Source hashes unchanged; test durations are diagnostic, not game-performance measurements. |
| 2026-11-16 | feat/microphone-test-and-gain 4240dd3ac2afd768b2d11f94896fad6a9fea230a plus uncommitted source-archive SHA-256 dc13c56616421df3c7e3dcb3f673d87b23f8a488b0186029151264a33408b3e9 | Native Ubuntu laptop; distro/kernel version not recorded for this run; no window | Not used; game-free fixtures | Not recorded for this fixture run | Python 3.14 (isolated environment); LLVM 23.1.6; fake audio/UI; tests/run_all.py, 4 jobs; MAGICBEAN_TOOLING set for retries/controls, unset in full initial run | Complete maintained game-free suite, 666 checks including skips; no gameplay | linux-full: 629 passed, 20 failed, 17 skipped in 303.93 s. Complete suite; two setup failures from an unset shared Tooling path cleared in the recorded retry. Remaining failures reproduce on pinned main. Source hashes unchanged; test durations are diagnostic, not game-performance measurements. |
| 2026-11-16 | feat/microphone-test-and-gain 4240dd3ac2afd768b2d11f94896fad6a9fea230a plus uncommitted source-archive SHA-256 dc13c56616421df3c7e3dcb3f673d87b23f8a488b0186029151264a33408b3e9 | Native Ubuntu laptop; distro/kernel version not recorded for this run; no window | Not used; game-free fixtures | Not recorded for this fixture run | Python 3.14 (isolated environment); LLVM 23.1.6; fake audio/UI; 4 concurrent fixtures; existing tools only | Focused 30-script mic/audio selection; no window, GPU or hardware capture | linux-focus: 30 passed, 0 failed, 0 skipped in 2.66 s. All mic/audio fixtures passed, including nine POSIX atomic-save cases. Source hashes unchanged; test durations are diagnostic, not game-performance measurements. |
| 2026-11-16 | bc0997d444f951445653a7d7aa49f78471a18a7e plus reviewed guide documentation | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; GNOME desktop | RTX 5090 Laptop GPU; NVIDIA 595.99.02 | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | Installed Clang 23 and isolated Python; four fixture jobs; normal OS skips; source hashes verified unchanged | final-full-suite; maintained tests/run_all.py with JUnit; game-free only | 631 passed, 19 failed, 17 skipped, 667 checks in 305.34 s. Eighteen failures match prior pinned-main controls. The additional dev-packs failure reproduced on main with inherited LONGSHOT_NO_PREPARE=1; removing that private environment default restores the focused test on both main and topic. No repository fix or assertion weakening. No game build or launch. |
| 2026-11-16 | bc0997d444f951445653a7d7aa49f78471a18a7e | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; GNOME desktop | RTX 5090 Laptop GPU; NVIDIA 595.99.02 | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | 2560x1600 reported window/swapchain after startup; Auto retains saved display choices; normal optimization; gain 100%; test stopped; frame generation off; renderer cap 60; V-Sync Auto/mailbox; host HOS_FPS_MAX=1000 is a separate ceiling | default-topic-02; stationary KakarikoVillage_Child from identical copied save; 45 s warm-up per side, 30 s settling and 60 s measurement; main/topic/main/topic | Native FPS median 60.50 from 30 log intervals; observed sample range 59.0-61.2 FPS. Matching scene, position, camera and window verified. This capped test is not proof of zero CPU cost. |
| 2026-11-16 | bc0997d444f951445653a7d7aa49f78471a18a7e | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; GNOME desktop | RTX 5090 Laptop GPU; NVIDIA 595.99.02 | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | 2560x1600 reported window/swapchain after startup; Auto retains saved display choices; normal optimization; gain 100%; test stopped; frame generation off; renderer cap 60; V-Sync Auto/mailbox; host HOS_FPS_MAX=1000 is a separate ceiling | default-topic-01; stationary KakarikoVillage_Child from identical copied save; 45 s warm-up per side, 30 s settling and 60 s measurement; main/topic/main/topic | Native FPS median 60.55 from 30 log intervals; observed sample range 58.9-61.3 FPS. Matching scene, position, camera and window verified. This capped test is not proof of zero CPU cost. |
| 2026-11-16 | f792212691eb3d49a37ff90b5690f0da20077a41 | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; GNOME desktop | RTX 5090 Laptop GPU; NVIDIA 595.99.02 | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | 2560x1600 reported window/swapchain after startup; Auto retains saved display choices; normal optimization; gain 100%; test stopped; frame generation off; renderer cap 60; V-Sync Auto/mailbox; host HOS_FPS_MAX=1000 is a separate ceiling | default-main-02; stationary KakarikoVillage_Child from identical copied save; 45 s warm-up per side, 30 s settling and 60 s measurement; main/topic/main/topic | Native FPS median 60.70 from 30 log intervals; observed sample range 60.2-62.2 FPS. Matching scene, position, camera and window verified. This capped test is not proof of zero CPU cost. |
| 2026-11-16 | f792212691eb3d49a37ff90b5690f0da20077a41 | Ubuntu 26.04.1 LTS; kernel 7.0.0-43-generic; GNOME desktop | RTX 5090 Laptop GPU; NVIDIA 595.99.02 | Core Ultra 9 275HX; about 61.2 GiB kernel-visible RAM | 2560x1600 reported window/swapchain after startup; Auto retains saved display choices; normal optimization; gain 100%; test stopped; frame generation off; renderer cap 60; V-Sync Auto/mailbox; host HOS_FPS_MAX=1000 is a separate ceiling | default-main-01; stationary KakarikoVillage_Child from identical copied save; 45 s warm-up per side, 30 s settling and 60 s measurement; main/topic/main/topic | Native FPS median 60.90 from 29 log intervals; observed sample range 59.0-61.9 FPS. Matching scene, position, camera and window verified. This capped test is not proof of zero CPU cost. |
| 2026-11-16 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x800, render resolution 100%, Quality preset Custom, frame generation off; each arm’s own translated program and pipeline caches (saved by one warm-up run of that arm) copied fresh, RX_VK_PRECOMPILE=0; no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT=soa | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; interleaved pairs, order aos, soa, soa, aos | Run fip-kN-soa-2: standing 9.56 fps, frame p50 99.3 / p99 153.5 ms, GPU time 72.8 ms; walking 10.42 fps / GPU 61.8 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-16 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x800, render resolution 100%, Quality preset Custom, frame generation off; each arm’s own translated program and pipeline caches (saved by one warm-up run of that arm) copied fresh, RX_VK_PRECOMPILE=0; no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT=soa | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; interleaved pairs, order aos, soa, soa, aos | Run fip-kN-soa-1: standing 9.34 fps, frame p50 101.4 / p99 150.5 ms, GPU time 72.6 ms; walking 9.15 fps / GPU 73.8 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-16 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x800, render resolution 100%, Quality preset Custom, frame generation off; each arm’s own translated program and pipeline caches (saved by one warm-up run of that arm) copied fresh, RX_VK_PRECOMPILE=0; no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT unset (aos) | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; interleaved pairs, order aos, soa, soa, aos | Run fip-kN-aos-2: standing 9.90 fps, frame p50 99.4 / p99 142.2 ms, GPU time 70.0 ms; walking 9.43 fps / GPU 71.4 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-16 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x800, render resolution 100%, Quality preset Custom, frame generation off; each arm’s own translated program and pipeline caches (saved by one warm-up run of that arm) copied fresh, RX_VK_PRECOMPILE=0; no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT unset (aos) | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; interleaved pairs, order aos, soa, soa, aos | Run fip-kN-aos-1: standing 8.96 fps, frame p50 105.9 / p99 161.3 ms, GPU time 77.1 ms; walking 9.45 fps / GPU 71.8 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-16 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; each arm’s own translated program and pipeline caches (saved by one warm-up run of that arm) copied fresh, RX_VK_PRECOMPILE=0; no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT=soa | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; interleaved pairs, order aos, soa, soa, aos | Run fip-kH-soa-2: standing 11.74 fps, frame p50 78.6 / p99 135.7 ms, GPU time 53.8 ms; walking 13.34 fps / GPU 41.1 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-16 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; each arm’s own translated program and pipeline caches (saved by one warm-up run of that arm) copied fresh, RX_VK_PRECOMPILE=0; no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT=soa | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; interleaved pairs, order aos, soa, soa, aos | Run fip-kH-soa-1: standing 11.58 fps, frame p50 79.6 / p99 135.8 ms, GPU time 52.0 ms; walking 12.73 fps / GPU 42.8 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-16 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; each arm’s own translated program and pipeline caches (saved by one warm-up run of that arm) copied fresh, RX_VK_PRECOMPILE=0; no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT unset (aos) | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; interleaved pairs, order aos, soa, soa, aos | Run fip-kH-aos-2: standing 11.53 fps, frame p50 81.5 / p99 132.6 ms, GPU time 54.9 ms; walking 12.01 fps / GPU 48.2 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-16 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; each arm’s own translated program and pipeline caches (saved by one warm-up run of that arm) copied fresh, RX_VK_PRECOMPILE=0; no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT unset (aos) | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; interleaved pairs, order aos, soa, soa, aos | Run fip-kH-aos-1: standing 10.71 fps, frame p50 88.0 / p99 142.7 ms, GPU time 60.1 ms; walking 11.14 fps / GPU 54.6 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-16 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; each arm’s own translated program and pipeline caches (saved by one warm-up run of that arm) copied fresh, RX_VK_PRECOMPILE=0; no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT=soa | the main outdoor area at the Lost Woods bridge, child, reached by teleport from the starting-area save, 60 s warm-up then 120 s standing; interleaved pairs, order aos, soa, soa, aos | Run fip-fH-soa-2: 22.68 fps, frame p50 44.1 / p99 48.9 ms, GPU time 22.8 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-16 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; each arm’s own translated program and pipeline caches (saved by one warm-up run of that arm) copied fresh, RX_VK_PRECOMPILE=0; no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT=soa | the main outdoor area at the Lost Woods bridge, child, reached by teleport from the starting-area save, 60 s warm-up then 120 s standing; interleaved pairs, order aos, soa, soa, aos | Run fip-fH-soa-1: 22.68 fps, frame p50 44.1 / p99 49.7 ms, GPU time 22.8 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-16 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; each arm’s own translated program and pipeline caches (saved by one warm-up run of that arm) copied fresh, RX_VK_PRECOMPILE=0; no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT unset (aos) | the main outdoor area at the Lost Woods bridge, child, reached by teleport from the starting-area save, 60 s warm-up then 120 s standing; interleaved pairs, order aos, soa, soa, aos | Run fip-fH-aos-2: 20.70 fps, frame p50 48.1 / p99 54.1 ms, GPU time 26.8 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-16 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; each arm’s own translated program and pipeline caches (saved by one warm-up run of that arm) copied fresh, RX_VK_PRECOMPILE=0; no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT unset (aos) | the main outdoor area at the Lost Woods bridge, child, reached by teleport from the starting-area save, 60 s warm-up then 120 s standing; interleaved pairs, order aos, soa, soa, aos | Run fip-fH-aos-1: 20.75 fps, frame p50 48.2 / p99 53.1 ms, GPU time 26.8 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-16 UTC | this PR’s gpu/ change (branch gpu-peak-power-state @ e4f7ed4b: sheep’s commit 16db7519a cherry-picked, and ours making it opt-in; the three new files byte-identical on the rebased head) as a renderer pack built via scripts/dev_packs.py on a base of main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed; pack longshot_vk.so sha256 d79401a56fab6f6ebb340857c297a4c9a390686c1280d8fb084b2e49ff867e62, the game binary main’s); the same pack in every run |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31); power_dpm_force_performance_level auto before every run |
Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared | 1280x720 output, render resolution 50% (option 4), the game’s own upscaler, frame generation off; precompile off (RX_VK_PRECOMPILE=0), the same warm caches copied fresh into every run, no longshot_env.txt; RX_VK_GPU_CLOCK=1 |
the starting forest, the child save’s spot outside the player character’s house (~2,200 draws a frame), 75 s warm-up then 45 s standing; one session 2026-11-16 03:30-04:04 UTC, interleaved off, on, on, off after a warm-up run, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy, APU power (power1_average) and temperature from sysfs at 10 Hz, the power level from sysfs every 2 s, read only |
run gc-on-kH-2, started 03:38:19 UTC, arm on: RX_VK_GPU_CLOCK=1 (log, once: vk: GPU clock: amdgpu's peak power state held while the game runs ... the power level auto -> profile_peak). 16.26 fps, 1% low 12.14 fps; present interval p50 / p95 / p99 60.5 / 72.4 / 80.3 ms, 0 over 100 ms; GPU time 31.8 ms a frame; GPU clock in the sample 1300 MHz 100%; GPU busy 58%; APU power mean / max 15.00 / 17.11 W; temperature max 72 C. Power level: profile_peak 03:38:22 (2 s after the game started) to 03:42:00; auto at 03:42:02 as the game exited (exit 03:42:03). Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | this PR’s gpu/ change (branch gpu-peak-power-state @ e4f7ed4b: sheep’s commit 16db7519a cherry-picked, and ours making it opt-in; the three new files byte-identical on the rebased head) as a renderer pack built via scripts/dev_packs.py on a base of main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed; pack longshot_vk.so sha256 d79401a56fab6f6ebb340857c297a4c9a390686c1280d8fb084b2e49ff867e62, the game binary main’s); the same pack in every run |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31); power_dpm_force_performance_level auto before every run |
Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared | 1280x720 output, render resolution 50% (option 4), the game’s own upscaler, frame generation off; precompile off (RX_VK_PRECOMPILE=0), the same warm caches copied fresh into every run, no longshot_env.txt; RX_VK_GPU_CLOCK=1 |
the starting forest, the child save’s spot outside the player character’s house (~2,200 draws a frame), 75 s warm-up then 45 s standing; one session 2026-11-16 03:30-04:04 UTC, interleaved off, on, on, off after a warm-up run, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy, APU power (power1_average) and temperature from sysfs at 10 Hz, the power level from sysfs every 2 s, read only |
run gc-on-kH-1, started 03:34:29 UTC, arm on: RX_VK_GPU_CLOCK=1 (log, once: vk: GPU clock: amdgpu's peak power state held while the game runs ... the power level auto -> profile_peak). 16.16 fps, 1% low 11.97 fps; present interval p50 / p95 / p99 61.1 / 73.4 / 80.9 ms, 0 over 100 ms; GPU time 32.0 ms a frame; GPU clock in the sample 1300 MHz 100%; GPU busy 58%; APU power mean / max 15.03 / 18.32 W; temperature max 72 C. Power level: profile_peak 03:34:29 (the first sample with the game) to 03:37:59; auto, 200 MHz, at 03:38:01 as the game exited (exit 03:38:01). Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | this PR’s gpu/ change (branch gpu-peak-power-state @ e4f7ed4b: sheep’s commit 16db7519a cherry-picked, and ours making it opt-in; the three new files byte-identical on the rebased head) as a renderer pack built via scripts/dev_packs.py on a base of main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed; pack longshot_vk.so sha256 d79401a56fab6f6ebb340857c297a4c9a390686c1280d8fb084b2e49ff867e62, the game binary main’s); the same pack in every run |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31); power_dpm_force_performance_level auto before every run |
Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared | native: 1280x800 output, render resolution 100%, the game’s own upscaler, frame generation off; precompile off (RX_VK_PRECOMPILE=0), the same warm caches copied fresh into every run, no longshot_env.txt; RX_VK_GPU_CLOCK=1 |
the main outdoor area (child, teleported from the starting-area save), 75 s warm-up then 45 s standing; one session 2026-11-16 03:30-04:04 UTC, interleaved off, on, on, off after a warm-up run, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy, APU power (power1_average) and temperature from sysfs at 10 Hz, the power level from sysfs every 2 s, read only |
run gc-on-fN-2, started 03:54:12 UTC, arm on: RX_VK_GPU_CLOCK=1 (log, once: vk: GPU clock: amdgpu's peak power state held while the game runs ... the power level auto -> profile_peak). 17.93 fps, 1% low 14.51 fps; present interval p50 / p95 / p99 55.5 / 59.2 / 61.7 ms, 1 over 100 ms; GPU time 35.6 ms a frame; GPU clock in the sample 1300 MHz 100%; GPU busy 70%; APU power mean / max 12.22 / 17.53 W; temperature max 72 C. Power level: profile_peak 03:54:14 to 03:57:50; auto at 03:57:52 as the game exited (exit 03:57:52). Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | this PR’s gpu/ change (branch gpu-peak-power-state @ e4f7ed4b: sheep’s commit 16db7519a cherry-picked, and ours making it opt-in; the three new files byte-identical on the rebased head) as a renderer pack built via scripts/dev_packs.py on a base of main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed; pack longshot_vk.so sha256 d79401a56fab6f6ebb340857c297a4c9a390686c1280d8fb084b2e49ff867e62, the game binary main’s); the same pack in every run |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31); power_dpm_force_performance_level auto before every run |
Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared | native: 1280x800 output, render resolution 100%, the game’s own upscaler, frame generation off; precompile off (RX_VK_PRECOMPILE=0), the same warm caches copied fresh into every run, no longshot_env.txt; RX_VK_GPU_CLOCK=1 |
the main outdoor area (child, teleported from the starting-area save), 75 s warm-up then 45 s standing; one session 2026-11-16 03:30-04:04 UTC, interleaved off, on, on, off after a warm-up run, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy, APU power (power1_average) and temperature from sysfs at 10 Hz, the power level from sysfs every 2 s, read only |
run gc-on-fN-1, started 03:50:15 UTC, arm on: RX_VK_GPU_CLOCK=1 (log, once: vk: GPU clock: amdgpu's peak power state held while the game runs ... the power level auto -> profile_peak). 17.83 fps, 1% low 14.98 fps; present interval p50 / p95 / p99 56.0 / 59.2 / 61.3 ms, 0 over 100 ms; GPU time 36.1 ms a frame; GPU clock in the sample 1300 MHz 100%; GPU busy 70%; APU power mean / max 12.11 / 16.48 W; temperature max 71 C. Power level: profile_peak 03:50:17 to 03:53:55; auto at 03:53:57 as the game exited (exit 03:53:59). Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | this PR’s gpu/ change (branch gpu-peak-power-state @ e4f7ed4b: sheep’s commit 16db7519a cherry-picked, and ours making it opt-in; the three new files byte-identical on the rebased head) as a renderer pack built via scripts/dev_packs.py on a base of main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed; pack longshot_vk.so sha256 d79401a56fab6f6ebb340857c297a4c9a390686c1280d8fb084b2e49ff867e62, the game binary main’s); the same pack in every run |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31); power_dpm_force_performance_level auto before every run |
Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared | 1280x720 output, render resolution 50% (option 4), the game’s own upscaler, frame generation off; precompile off (RX_VK_PRECOMPILE=0), the same warm caches copied fresh into every run, no longshot_env.txt; RX_VK_GPU_CLOCK unset |
the starting forest, the child save’s spot outside the player character’s house (~2,200 draws a frame), 75 s warm-up then 45 s standing; one session 2026-11-16 03:30-04:04 UTC, interleaved off, on, on, off after a warm-up run, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy, APU power (power1_average) and temperature from sysfs at 10 Hz, the power level from sysfs every 2 s, read only |
run gc-off-kH-2, started 03:42:16 UTC, arm off: RX_VK_GPU_CLOCK unset (main’s path: nothing opened or asked, no vk: GPU clock line). 11.39 fps, 1% low 6.86 fps; present interval p50 / p95 / p99 80.1 / 125.9 / 139.8 ms, 131 over 100 ms; GPU time 56.55 ms a frame; GPU clock in the sample 1040 MHz 33%, 200 MHz 25%, between 31%; GPU busy 78%; APU power mean / max 14.81 / 16.90 W; temperature max 69 C. Power level: auto throughout. Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | this PR’s gpu/ change (branch gpu-peak-power-state @ e4f7ed4b: sheep’s commit 16db7519a cherry-picked, and ours making it opt-in; the three new files byte-identical on the rebased head) as a renderer pack built via scripts/dev_packs.py on a base of main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed; pack longshot_vk.so sha256 d79401a56fab6f6ebb340857c297a4c9a390686c1280d8fb084b2e49ff867e62, the game binary main’s); the same pack in every run |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31); power_dpm_force_performance_level auto before every run |
Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared | 1280x720 output, render resolution 50% (option 4), the game’s own upscaler, frame generation off; precompile off (RX_VK_PRECOMPILE=0), the same warm caches copied fresh into every run, no longshot_env.txt; RX_VK_GPU_CLOCK unset |
the starting forest, the child save’s spot outside the player character’s house (~2,200 draws a frame), 75 s warm-up then 45 s standing; one session 2026-11-16 03:30-04:04 UTC, interleaved off, on, on, off after a warm-up run, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy, APU power (power1_average) and temperature from sysfs at 10 Hz, the power level from sysfs every 2 s, read only |
run gc-off-kH-1, started 03:30:32 UTC, arm off: RX_VK_GPU_CLOCK unset (main’s path: nothing opened or asked, no vk: GPU clock line). 10.85 fps, 1% low 6.68 fps; present interval p50 / p95 / p99 86.6 / 131.1 / 141.8 ms, 170 over 100 ms; GPU time 59.7 ms a frame; GPU clock in the sample 1040 MHz 30%, 200 MHz 29%, between 32%; GPU busy 80%; APU power mean / max 14.87 / 17.41 W; temperature max 69 C. Power level: auto throughout. Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | this PR’s gpu/ change (branch gpu-peak-power-state @ e4f7ed4b: sheep’s commit 16db7519a cherry-picked, and ours making it opt-in; the three new files byte-identical on the rebased head) as a renderer pack built via scripts/dev_packs.py on a base of main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed; pack longshot_vk.so sha256 d79401a56fab6f6ebb340857c297a4c9a390686c1280d8fb084b2e49ff867e62, the game binary main’s); the same pack in every run |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31); power_dpm_force_performance_level auto before every run |
Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared | native: 1280x800 output, render resolution 100%, the game’s own upscaler, frame generation off; precompile off (RX_VK_PRECOMPILE=0), the same warm caches copied fresh into every run, no longshot_env.txt; RX_VK_GPU_CLOCK unset |
the main outdoor area (child, teleported from the starting-area save), 75 s warm-up then 45 s standing; one session 2026-11-16 03:30-04:04 UTC, interleaved off, on, on, off after a warm-up run, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy, APU power (power1_average) and temperature from sysfs at 10 Hz, the power level from sysfs every 2 s, read only |
run gc-off-fN-2, started 03:58:05 UTC, arm off: RX_VK_GPU_CLOCK unset (main’s path: nothing opened or asked, no vk: GPU clock line). 14.95 fps, 1% low 13.60 fps; present interval p50 / p95 / p99 66.7 / 70.3 / 72.6 ms, 0 over 100 ms; GPU time 45.1 ms a frame; GPU clock in the sample 1040 MHz 88%, 200 MHz 12%; GPU busy 78%; APU power mean / max 8.11 / 11.57 W; temperature max 63 C. Power level: auto throughout. Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | this PR’s gpu/ change (branch gpu-peak-power-state @ e4f7ed4b: sheep’s commit 16db7519a cherry-picked, and ours making it opt-in; the three new files byte-identical on the rebased head) as a renderer pack built via scripts/dev_packs.py on a base of main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed; pack longshot_vk.so sha256 d79401a56fab6f6ebb340857c297a4c9a390686c1280d8fb084b2e49ff867e62, the game binary main’s); the same pack in every run |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31); power_dpm_force_performance_level auto before every run |
Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared | native: 1280x800 output, render resolution 100%, the game’s own upscaler, frame generation off; precompile off (RX_VK_PRECOMPILE=0), the same warm caches copied fresh into every run, no longshot_env.txt; RX_VK_GPU_CLOCK unset |
the main outdoor area (child, teleported from the starting-area save), 75 s warm-up then 45 s standing; one session 2026-11-16 03:30-04:04 UTC, interleaved off, on, on, off after a warm-up run, for each scene; fps and frame times from HOS_PRESENT_TRACE in the sample (1% low: 1000 / the mean of the slowest 1% of present intervals), GPU time from the log, GPU clock, busy, APU power (power1_average) and temperature from sysfs at 10 Hz, the power level from sysfs every 2 s, read only |
run gc-off-fN-1, started 03:46:11 UTC, arm off: RX_VK_GPU_CLOCK unset (main’s path: nothing opened or asked, no vk: GPU clock line). 15.01 fps, 1% low 12.77 fps; present interval p50 / p95 / p99 66.4 / 70.4 / 74.2 ms, 0 over 100 ms; GPU time 44.9 ms a frame; GPU clock in the sample 1040 MHz 88%, 200 MHz 12%; GPU busy 70%; APU power mean / max 8.17 / 13.06 W; temperature max 63 C. Power level: auto throughout. Exit 0, no device fault, no VK_ERROR, no amdgpu kernel line |
| 2026-11-16 UTC | this PR’s gpu/ change (branch gpu-peak-power-state @ e4f7ed4b: sheep’s commit 16db7519a cherry-picked, and ours making it opt-in; the three new files byte-identical on the rebased head) as a renderer pack built via scripts/dev_packs.py on a base of main 9c9d69ed (longshot v3.0.0-rc.3-75-g9c9d69ed; pack longshot_vk.so sha256 d79401a56fab6f6ebb340857c297a4c9a390686c1280d8fb084b2e49ff867e62, the game binary main’s); the same pack in every run |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2), headless (HOS_HEADLESS=1), plugged in |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31); power_dpm_force_performance_level auto before every run |
Steam Deck OLED APU (Zen 2, 4 cores / 8 threads), 16 GB shared | 1280x720 output, render resolution 50% (option 4), the game’s own upscaler, frame generation off; precompile off (RX_VK_PRECOMPILE=0), the same warm caches copied fresh into every run, no longshot_env.txt; RX_VK_GPU_CLOCK=1 |
starting-area child save from start-up, the game killed with kill -9 about 30 s after the vk: GPU clock line; the same session as the timed runs, after the last of them; the power level and the clock read from sysfs before, during and after, read only |
run gc-kill-1, started 04:02:38 UTC. 04:02:33, before: auto. The game logged vk: GPU clock: amdgpu's peak power state held while the game runs ... auto -> profile_peak; 04:03:08: profile_peak, 1300 MHz; kill -9 of the game’s process (exit -9); 04:03:09, right after: auto, 200 MHz; still auto at 5 s and 20 s after. No kernel line about amdgpu or drm in the journal for those minutes |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 with the render-thread changes since merged to main (#761, #838, 64ed6a2f) and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), fix/vk-capture-collect-window’s three commits (its diff as at fd5238f7) and a local counter of collect_mu’s waits (in no pull request) |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | RX_VK_FRAME_GATE=0, RX_VK_PRESENT_AFTER_SIGNAL=1; HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (Ultra Quality), frame generation off |
the starting forest (child), the Market (child), the starting forest, then the Market in one session by the debug teleport (four area loads), standing at each arrival spot: 5 s to settle, then 20 s measured in each; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame of the measured windows (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames), the capture’s waits from the local counter over the whole session, the boot and the loads included |
the starting forest 89.7 and 87.7 fps, the Market 57.4 and 57.5 fps (medians of the 5 s windows); 0 of 5,400 frames with a dark patch (the largest change 1.5); a capture waited for the collection 33,267 times in the session, 2,514 ms in all, the longest 20.7 ms (at an area load); no device loss (run LK0-2) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 with the render-thread changes since merged to main (#761, #838, 64ed6a2f) and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), fix/vk-capture-collect-window’s three commits (its diff as at fd5238f7) and a local counter of collect_mu’s waits (in no pull request) |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | Whole frames on (the default); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (Ultra Quality), frame generation off |
the starting forest (child), the Market (child), the starting forest, then the Market in one session by the debug teleport (four area loads), standing at each arrival spot: 5 s to settle, then 20 s measured in each; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame of the measured windows (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames), the capture’s waits from the local counter over the whole session, the boot and the loads included |
the starting forest 83.0 and 81.3 fps, the Market 53.9 and 51.7 fps (medians of the 5 s windows); 0 of 5,400 frames with a dark patch (the largest change 1.0); a capture waited for the collection 932 times in the session, 323 ms in all, the longest 11.9 ms (at an area load); no device loss (run LK1-2) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 with the render-thread changes since merged to main (#761, #838, 64ed6a2f) and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), fix/vk-capture-collect-window’s three commits (its diff as at fd5238f7) and a local counter of collect_mu’s waits (in no pull request) |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | RX_VK_FRAME_GATE=0, RX_VK_PRESENT_AFTER_SIGNAL=1; HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (Ultra Quality), frame generation off |
the starting forest (child), the Market (child), the starting forest, then the Market in one session by the debug teleport (four area loads), standing at each arrival spot: 5 s to settle, then 20 s measured in each; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame of the measured windows (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames), the capture’s waits from the local counter over the whole session, the boot and the loads included |
the starting forest 88.6 and 89.8 fps, the Market 56.3 and 57.2 fps (medians of the 5 s windows); 0 of 5,400 frames with a dark patch (the largest change 2.4); a capture waited for the collection 35,165 times in the session, 2,653 ms in all, the longest 20.8 ms (at an area load); no device loss (run LK0) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 with the render-thread changes since merged to main (#761, #838, 64ed6a2f) and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), fix/vk-capture-collect-window’s three commits (its diff as at fd5238f7) and a local counter of collect_mu’s waits (in no pull request) |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | Whole frames on (the default); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured windows (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (Ultra Quality), frame generation off |
the starting forest (child), the Market (child), the starting forest, then the Market in one session by the debug teleport (four area loads), standing at each arrival spot: 5 s to settle, then 20 s measured in each; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame of the measured windows (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames), the capture’s waits from the local counter over the whole session, the boot and the loads included |
the starting forest 83.1 and 77.8 fps, the Market 51.7 and 54.0 fps (medians of the 5 s windows); 0 of 5,400 frames with a dark patch (the largest change 2.0); a capture waited for the collection 831 times in the session, 248 ms in all, the longest 0.99 ms; no device loss (run LK1) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 with the render-thread changes since merged to main (#761, #838, 64ed6a2f) and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and fix/vk-capture-collect-window’s three commits (its diff as at fd5238f7) |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | RX_VK_FRAME_GATE=0, RX_VK_PRESENT_AFTER_SIGNAL=1; HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured window (RX_VK_FRAMELOG), its pictures kept for scripts/framelog.py and deleted after; 1280x720 window, rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (Ultra Quality), frame generation off |
the starting forest (child), standing at the debug teleport’s arrival spot: 40 s to settle, then 40 s measured; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames) |
94.0 fps (median of 8 windows); 0 of 3,600 frames with a dark patch (the largest change 1.0); scripts/framelog.py: 0 events in 3,785 frames; no device loss (run RBK, after) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 with the render-thread changes since merged to main (#761, #838, 64ed6a2f) and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0); before fix/vk-capture-collect-window |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | RX_VK_FRAME_GATE=0, RX_VK_PRESENT_AFTER_SIGNAL=1; HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured window (RX_VK_FRAMELOG), its pictures kept for scripts/framelog.py and deleted after; 1280x720 window, rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (Ultra Quality), frame generation off |
the starting forest (child), standing at the debug teleport’s arrival spot: 40 s to settle, then 40 s measured; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames) |
92.0 fps (median of 8 windows); 27 of 3,600 frames with a dark patch (the largest 28.4 levels darker) and 1 with a bright one; scripts/framelog.py: 57 events in 3,715 frames; no device loss (run RAK, before) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 with the render-thread changes since merged to main (#761, #838, 64ed6a2f) and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0); before fix/vk-capture-collect-window |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | RX_VK_FRAME_GATE=0, RX_VK_PRESENT_AFTER_SIGNAL=1; HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured window (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (Ultra Quality), frame generation off |
the starting forest (child), standing at the debug teleport’s arrival spot: 40 s to settle, then 40 s measured; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames) |
93.1 fps (median of 8 windows); 16 of 3,600 frames with a dark patch (the largest 28.9 levels darker); no device loss (run RA0-2, before) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 with the render-thread changes since merged to main (#761, #838, 64ed6a2f) and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and fix/vk-capture-collect-window’s three commits (its diff as at fd5238f7) |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | RX_VK_FRAME_GATE=0, RX_VK_PRESENT_AFTER_SIGNAL=1; HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured window (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (Ultra Quality), frame generation off |
the starting forest (child), standing at the debug teleport’s arrival spot: 40 s to settle, then 40 s measured; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames) |
92.5 fps (median of 8 windows); 0 of 3,600 frames with a dark patch (the largest change 0.4); no device loss (run RB0-2, after) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 with the render-thread changes since merged to main (#761, #838, 64ed6a2f) and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0); before fix/vk-capture-collect-window |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | Whole frames on (the default); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured window (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (Ultra Quality), frame generation off |
the starting forest (child), standing at the debug teleport’s arrival spot: 40 s to settle, then 40 s measured; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames) |
86.5 fps (median of 8 windows); 0 of 3,000 frames with a dark patch; no device loss (run RA1-2, before) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 with the render-thread changes since merged to main (#761, #838, 64ed6a2f) and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and fix/vk-capture-collect-window’s three commits (its diff as at fd5238f7) |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | Whole frames on (the default); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured window (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (Ultra Quality), frame generation off |
the starting forest (child), standing at the debug teleport’s arrival spot: 40 s to settle, then 40 s measured; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames) |
86.6 fps (median of 8 windows); 0 of 3,000 frames with a dark patch; no device loss (run RB1-2, after) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 with the render-thread changes since merged to main (#761, #838, 64ed6a2f) and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and fix/vk-capture-collect-window’s three commits (its diff as at fd5238f7) |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | RX_VK_FRAME_GATE=0, RX_VK_PRESENT_AFTER_SIGNAL=1; HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured window (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (Ultra Quality), frame generation off |
the starting forest (child), standing at the debug teleport’s arrival spot: 40 s to settle, then 40 s measured; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames) |
93.4 fps (median of 8 windows); 0 of 3,600 frames with a dark patch (the largest change 0.9); no device loss (run RB0, after) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 with the render-thread changes since merged to main (#761, #838, 64ed6a2f) and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0); before fix/vk-capture-collect-window |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | RX_VK_FRAME_GATE=0, RX_VK_PRESENT_AFTER_SIGNAL=1; HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured window (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (Ultra Quality), frame generation off |
the starting forest (child), standing at the debug teleport’s arrival spot: 40 s to settle, then 40 s measured; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames) |
92.6 fps (median of 8 windows); 17 of 3,600 frames with a dark patch (the largest 31.0 levels darker than both neighbours) and 1 with a bright one; no device loss (run RA0, before) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 with the render-thread changes since merged to main (#761, #838, 64ed6a2f) and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0), and fix/vk-capture-collect-window’s three commits (its diff as at fd5238f7) |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | Whole frames on (the default); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured window (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (Ultra Quality), frame generation off |
the starting forest (child), standing at the debug teleport’s arrival spot: 40 s to settle, then 40 s measured; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames) |
86.4 fps (median of 8 windows); 0 of 3,000 frames with a dark patch; no device loss (run RB1, after) |
| 2026-11-16 | the renderer built alone (RX_VK_ONLY, llvm-mingw) from 690d8446 with the render-thread changes since merged to main (#761, #838, 64ed6a2f) and a local region test (in no pull request), in a nightly-80-7315957c game folder (host 690d8446) never prepared (RX_VK_PRECOMPILE=0); before fix/vk-capture-collect-window |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | Whole frames on (the default); HOS_PROFILE=5, HOS_FPS_MAX=240; the frame log on over the measured window (RX_VK_FRAMELOG), its pictures dropped (RX_VK_FRAMELOG_DROP=1); 1280x720 window, rendered at 960x540 and upscaled with Intel XeSS 2.0.2 (Ultra Quality), frame generation off |
the starting forest (child), standing at the debug teleport’s arrival spot: 40 s to settle, then 40 s measured; fps from the log’s 5 s vi: windows, dark frames from the region test on every presented frame (2 or more of 16x9 regions more than 10 levels darker than in both neighbouring frames) |
86.5 fps (median of 8 windows); 0 of 3,000 frames with a dark patch; no device loss (run RA1, before) |
| 2026-11-16 | longshot v3.0.0-rc.5 on main 71b1edd8 with this change in the working tree (runs check-warm2, check-cold); the cold-1 run before it on main 13e90a5e with the VRAM changes of #1158 beside it; built with make_game.py |
Windows 11 Pro 26200, no window (headless, scripted input) | RTX 4070 SUPER 12 GB, NVIDIA 617.42 | Ryzen 9 9950X3D (16 cores, SMT off), 32 GB | 2560x1440, DLAA, no frame generation, uncapped (HOS_FPS_MAX=1000); RX_VK_LOAD=create, RX_VK_TEST_ALLOWED_MB=10200; a cold driver cache by __GL_SHADER_DISK_CACHE_PATH=<an empty folder> (honoured by NVIDIA’s Windows driver) |
Castle Town’s gate into the main outdoor area and back: save set “First Step into the game world Town” (gate-bridge-wall.txt, wall clock); one run each | Warm driver cache: has 48 of 48 noted pipelines asked (a sample, in 224 ms), nothing built again. Cold driver cache before this change (cold-1): 10,263 listed pipelines in no cache, 702,312 draws left out in 21 of 163 two-second spans, 21.3 s of draw waits. Cold with it (check-cold): has 0 of 48 ... (in 87 ms): it lost them, 39,171 noted pipelines and 9,927 compute pipelines built before the title in 271.2 s on 16 threads, then 456 draws left out in 5 of 191 spans and no waits |
| 2026-11-15 | the all-zero check change’s gpu/vk diff (f83df428) applied on main 9c9d69ed, renderer-only pack zholes, with RX_VK_ZERO_HOLES=0 | SteamOS 3.9.5 (kernel 7.2.7-valve2), headless over ssh | AMD Custom GPU 0932 (Steam Deck OLED), RADV Mesa 26.2.0-devel (git-5c65d4fc31), stock clock | Steam Deck OLED APU, 16 GB (8 GB swap) | 1280x800 output, render resolution 100%, Quality preset Custom, frame generation off; code defaults except RX_VK_PRECOMPILE=0 with warm caches | the main outdoor area from a child starting-area save (teleport to the Lost Woods bridge, which loads the 288 MB shadow map): 45 s standing, 3 camera-pan cycles, standing to 180 s after the arrival; interleaved main, PR, PR with RX_VK_ZERO_HOLES=0, then the reverse (runs this file is run zh-off-2 of six) | zh-off-2: the 288 MB shadow map’s all-zero check 310.1 ms (288 MB read); peak guest memory backing 5892 MiB; 14.42 fps standing, 12.99 fps in the pans (118 frames over 100 ms); lowest MemAvailable 810 MiB; exited normally |
| 2026-11-15 | the all-zero check change’s gpu/vk diff (f83df428) applied on main 9c9d69ed, renderer-only pack zholes, with RX_VK_ZERO_HOLES=0 | SteamOS 3.9.5 (kernel 7.2.7-valve2), headless over ssh | AMD Custom GPU 0932 (Steam Deck OLED), RADV Mesa 26.2.0-devel (git-5c65d4fc31), stock clock | Steam Deck OLED APU, 16 GB (8 GB swap) | 1280x800 output, render resolution 100%, Quality preset Custom, frame generation off; code defaults except RX_VK_PRECOMPILE=0 with warm caches | the main outdoor area from a child starting-area save (teleport to the Lost Woods bridge, which loads the 288 MB shadow map): 45 s standing, 3 camera-pan cycles, standing to 180 s after the arrival; interleaved main, PR, PR with RX_VK_ZERO_HOLES=0, then the reverse (runs this file is run zh-off-1 of six) | zh-off-1: the 288 MB shadow map’s all-zero check 300.1 ms (288 MB read); peak guest memory backing 5880 MiB; 14.46 fps standing, 13.45 fps in the pans (75 frames over 100 ms); lowest MemAvailable 1150 MiB; exited normally |
| 2026-11-15 | main 9c9d69ed (renderer-only dev_packs base, no pack) | SteamOS 3.9.5 (kernel 7.2.7-valve2), headless over ssh | AMD Custom GPU 0932 (Steam Deck OLED), RADV Mesa 26.2.0-devel (git-5c65d4fc31), stock clock | Steam Deck OLED APU, 16 GB (8 GB swap) | 1280x800 output, render resolution 100%, Quality preset Custom, frame generation off; code defaults except RX_VK_PRECOMPILE=0 with warm caches | the main outdoor area from a child starting-area save (teleport to the Lost Woods bridge, which loads the 288 MB shadow map): 45 s standing, 3 camera-pan cycles, standing to 180 s after the arrival; interleaved main, PR, PR with RX_VK_ZERO_HOLES=0, then the reverse (runs this file is run zh-main-2 of six) | zh-main-2: the 288 MB shadow map’s all-zero check 321.6 ms (whole storage read); peak guest memory backing 5914 MiB; 14.53 fps standing, 13.33 fps in the pans (97 frames over 100 ms); lowest MemAvailable 882 MiB; exited normally |
| 2026-11-15 | main 9c9d69ed (renderer-only dev_packs base, no pack) | SteamOS 3.9.5 (kernel 7.2.7-valve2), headless over ssh | AMD Custom GPU 0932 (Steam Deck OLED), RADV Mesa 26.2.0-devel (git-5c65d4fc31), stock clock | Steam Deck OLED APU, 16 GB (8 GB swap) | 1280x800 output, render resolution 100%, Quality preset Custom, frame generation off; code defaults except RX_VK_PRECOMPILE=0 with warm caches | the main outdoor area from a child starting-area save (teleport to the Lost Woods bridge, which loads the 288 MB shadow map): 45 s standing, 3 camera-pan cycles, standing to 180 s after the arrival; interleaved main, PR, PR with RX_VK_ZERO_HOLES=0, then the reverse (runs this file is run zh-main-1 of six) | zh-main-1: the 288 MB shadow map’s all-zero check 299.9 ms (whole storage read); peak guest memory backing 5911 MiB; 14.49 fps standing, 13.38 fps in the pans (100 frames over 100 ms); lowest MemAvailable 732 MiB; exited normally |
| 2026-11-15 | the all-zero check change’s gpu/vk diff (f83df428) applied on main 9c9d69ed, renderer-only pack zholes | SteamOS 3.9.5 (kernel 7.2.7-valve2), headless over ssh | AMD Custom GPU 0932 (Steam Deck OLED), RADV Mesa 26.2.0-devel (git-5c65d4fc31), stock clock | Steam Deck OLED APU, 16 GB (8 GB swap) | 1280x800 output, render resolution 100%, Quality preset Custom, frame generation off; code defaults except RX_VK_PRECOMPILE=0 with warm caches | the main outdoor area from a child starting-area save (teleport to the Lost Woods bridge, which loads the 288 MB shadow map): 45 s standing, 3 camera-pan cycles, standing to 180 s after the arrival; interleaved main, PR, PR with RX_VK_ZERO_HOLES=0, then the reverse (runs this file is run zh-def-2 of six) | zh-def-2: the 288 MB shadow map’s all-zero check 6.0 ms (16 MB read, the rest holes); peak guest memory backing 5626 MiB; 14.49 fps standing, 13.72 fps in the pans (68 frames over 100 ms); lowest MemAvailable 1189 MiB; picture as main’s; exited normally |
| 2026-11-15 | the all-zero check change’s gpu/vk diff (f83df428) applied on main 9c9d69ed, renderer-only pack zholes | SteamOS 3.9.5 (kernel 7.2.7-valve2), headless over ssh | AMD Custom GPU 0932 (Steam Deck OLED), RADV Mesa 26.2.0-devel (git-5c65d4fc31), stock clock | Steam Deck OLED APU, 16 GB (8 GB swap) | 1280x800 output, render resolution 100%, Quality preset Custom, frame generation off; code defaults except RX_VK_PRECOMPILE=0 with warm caches | the main outdoor area from a child starting-area save (teleport to the Lost Woods bridge, which loads the 288 MB shadow map): 45 s standing, 3 camera-pan cycles, standing to 180 s after the arrival; interleaved main, PR, PR with RX_VK_ZERO_HOLES=0, then the reverse (runs this file is run zh-def-1 of six) | zh-def-1: the 288 MB shadow map’s all-zero check 6.2 ms (16 MB read, the rest holes); peak guest memory backing 5650 MiB; 14.50 fps standing, 13.44 fps in the pans (88 frames over 100 ms); lowest MemAvailable 1192 MiB; picture as main’s; exited normally |
| 2026-11-15 | 05e45da9, PR #614 | Windows 11 build 26200 | RTX 4090 / Vulkan driver ID 0x9b438000 |
Ryzen 7 7700X / 32 GB | 1920x1080 offscreen, Debug tools Off, Transition autosave On, 75-second wait after staging | Castle Town Market exit into the main outdoor area | Codex run08. One progress-save write and all four transition requests allowed. |
| 2026-11-15 | 05e45da9, PR #614 | Windows 11 build 26200 | RTX 4090 / Vulkan driver ID 0x9b438000 |
Ryzen 7 7700X / 32 GB | 1920x1080 offscreen, Debug tools Off, Transition autosave Off, 75-second wait after staging | Castle Town Market exit into the main outdoor area | Codex run07. Zero progress-save writes and all four transition requests blocked. |
| 2026-11-15 | c2e30565, PR #614 | Windows 11 build 26200 | RTX 4090 / Vulkan driver ID 0x9b438000 |
Ryzen 7 7700X / 32 GB | 1920x1080, Debug tools Off, Transition autosave Off, 75-second wait after staging | Gerudo Valley exit into the main outdoor area | Codex run06: The walking retry completed, with zero progress-save writes and all four transition requests blocked. |
| 2026-11-15 | c2e30565, PR #614 | Windows 11 build 26200 | RTX 4090 / Vulkan driver ID 0x9b438000 |
Ryzen 7 7700X / 32 GB | 1920x1080, Debug tools Off, Transition autosave On, 75-second wait after staging | Gerudo Valley, Market and the starting forest exits into the main outdoor area | Codex run05: One progress-save write per walk, three total. Each crossing logged the same four transition requests, with zero suppressed. |
| 2026-11-15 | c2e30565, PR #614 | Windows 11 build 26200 | RTX 4090 / Vulkan driver ID 0x9b438000 |
Ryzen 7 7700X / 32 GB | 1920x1080, Debug tools Off, Transition autosave Off, 75-second wait after staging | Market and the starting forest exits into the main outdoor area | Codex run04: Zero progress-save writes on either walk. Four transition requests blocked per crossing. Gerudo staging timed out and was retried separately. |
| 2026-11-15 | 8fd316cd, PR #614 | Windows 11 build 26200 | RTX 4090 / Vulkan driver ID 0x9b438000 |
Ryzen 7 7700X / 32 GB | 2560x1440, Debug tools Off, Transition autosave On, save trace enabled | the main outdoor area → the village → the main outdoor area, waiting 75 seconds before each crossing | Codex run03: One new progress save per crossing, two total. No requests were suppressed. |
| 2026-11-15 | 8fd316cd, PR #614 | Windows 11 build 26200 | RTX 4090 / Vulkan driver ID 0x9b438000 |
Ryzen 7 7700X / 32 GB | 2560x1440, Debug tools Off, Transition autosave Off, save trace enabled | the main outdoor area → the village → the main outdoor area, waiting 75 seconds before each crossing | Codex run02: Both crossings left the progress save unchanged, with 10 suppressed requests including startup. System → Save then changed one slot. |
| 2026-11-15 | 3b54a58d, PR #614 | Windows 11 build 26200 | RTX 4090 / Vulkan driver ID 0x9b438000 |
Ryzen 7 7700X / 32 GB | 2560x1440, Debug tools Off, Transition autosave Off then On, local save trace enabled | Walked between the main outdoor area and the village, then used System → Save | Codex run01: Off persisted across restart. Four transition requests were blocked, but two new progress saves still appeared, one per crossing. A manual save changed one slot with Off. With On, a repeat into the village wrote a new slot with no transition suppression. |
| 2026-11-15 UTC | Longshot main a4821d56 (includes #488 pacing; the pacing string is in the host binary; runtime as run01) | SteamOS holo (BUILD_ID 20260922.1), KDE Desktop mode, unattended scripted input through the game’s own control bridge | AMD Custom GPU 0405 (RADV VANGOGH), Steam Deck LCD (Jupiter), 8 GB shared VRAM, Mesa 26.1.2, Vulkan 1.4, driver 0x6801063 | AMD Vangogh 4c/8t APU, 16 GB LPDDR5; runtime and ~9 GB pipeline cache on a UHS-I SD card | RX_VK_READBACK_LATE=newest HOS_SHORT_SLEEP_TIMER=1 RX_VK_MESH_OPT=1 RX_VK_BC_LAYERS=8 RX_VK_TEX_MAX_DIM=512 RX_VK_TEX_SKIP_MB=8; game settings as run01 (Auto/Low/FSR 3.1) but the test ini’s frame rate cap row reads 144; GPU level auto, no clock pin (the plain SteamOS state a Desktop-mode launch starts in) |
the opening movie: boot, let the movie play on its own, read a movie of 1800 pictures on libdav1d: … N skipped from the log at movie end; two cold/warm arms |
run04 (stock clocks): arm 1: movie one 30 skipped, movie two 256; arm 2: 31 and 94 — each of 1,800, decode 2.0-4.9 ms/picture, one shown every 16.7 ms. For scale: issue #73’s pre-fix Deck evidence recorded 1,055 of 5,947 pictures skipped across the whole intro at the 60 cap (settings differed, so this is not a #488-alone A/B). Note: the pacing engage line (vi: a movie is playing…) did not fire in these arms; why is not established here. |
| 2026-11-15 UTC | Longshot main aa95d35f (includes #735 per-card VRAM budgets and #761 Arc occlusion copy; runtime packed on the Deck with dev_packs --no-prepare, object cache warm) |
SteamOS holo (BUILD_ID 20260922.1), KDE Desktop mode, unattended scripted input through the game’s own control bridge | AMD Custom GPU 0405 (RADV VANGOGH), Steam Deck LCD (Jupiter), 8 GB shared VRAM, Mesa 26.1.2, Vulkan 1.4, driver 0x6801063 | AMD Vangogh 4c/8t APU, 16 GB LPDDR5; runtime and ~9 GB pipeline cache on a UHS-I SD card (the worse case, see #386) | Auto (this PC): Low preset, FSR 3.1 at the game’s resolution, 60 fps cap, frame pacing on, threaded off; RX_VK_READBACK_LATE=newest HOS_SHORT_SLEEP_TIMER=1 RX_VK_MESH_OPT=1 RX_VK_BC_LAYERS=8 RX_VK_TEX_MAX_DIM=512 RX_VK_TEX_SKIP_MB=8; GPU level high (active sclk 2:1600Mhz), CPU governors performance, set before each arm and restored after |
deck-spin v4 protocol, one boot per arm: load save, settle 45 s, walk forward 10 s into the scene, settle 20 s, scripted idle x10, camera spin x6. Two scenes: the starting-area Village benchmark save (the worst steady outdoor save we have) and the the game world Town entrance save | run03: starting-area idle 30.9 [29.5-31.5], spin 27.6 [22.6-31.4]; the game world Town idle 26.0 [25.5-26.5], spin 20.4 [16.5-23.4]. Warm-up boot ran first on the same runtime; the worst frame of the arm’s 600 was 272 ms at starting-area and the log noted 19 re-upload events. Same save, older main a4821d56 with the same switches and pin, another runtime: idle 24-25.6 (arms posted on #716) — main’s #735 and #761 work is the likeliest reason starting-area improved, not proven alone. |
| 2026-11-15 | feat/sun-shadow-resolution e257da53 (the Shadow resolution setting, before it followed the Quality presets) as a pack over a main base with DLSS, reporting v3.0.0-rc.4-13-ge257da53; tester jiro113 | Windows 11 Home 10.0.26300; the game’s window shown | RTX 4090 24 GB, NVIDIA 621.24 | Core i9-13900K, 64 GB RAM | Output 2560x1440 with ultrawide on, presented 5120x1440 SDR; DLSS on, 2560x720 to 5120x1440, preset K; frame generation Off; V-Sync Auto (mailbox); Quality preset Auto (this PC), on its Ultra tier (memory_budget 5); RX_VK_PASS_TIMING=1; Shadow resolution switched live through the control pipe | the main outdoor area at the castle’s drawbridge as a child (position -0.04, 150.61, 32.97), standing still, camera untouched; in-session A/B, 3 rounds of Game’s, High, Highest, each setting 2 blocks of 600 presents after one block left out at the switch | Run a3. GPU time a frame, all passes (mean of 6 blocks, sd): Game’s 1600 9.03 ms (0.03), High 2200 9.06 ms (0.01), Highest 3072 9.10 ms (0.02); the main scene pass 3.558 / 3.587 / 3.626 ms; render thread 12.65 / 12.61 / 12.70 ms a frame, so the frame rate did not move (CPU-bound here). The pass timing shows the cascade passes drawn into 1600x1600, 2200x2200 and 3072x3072 depth targets for the three settings, and by eye the drawbridge chains’ shadows on the planks lose their stair steps on High and Highest. The sun’s shadow maps take about 31 / 58 / 113 MB of video memory, computed from their size (not measured). DLSS started (NGX), no device lost. |
| 2026-11-15 | main 7ff5256e against the combined build of main 7ff5256e with https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/550 to https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/558 merged (https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/557 @ f7f5578a), x86-64-v3, release play profile, LLVM 23.1.5 | SteamOS 3.9.5 (build 20261023.100), kernel 7.2.7-valve2, headless (HOS_HEADLESS=1), stderr on the journal |
AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel git-5c65d4fc31 | Steam Deck OLED, AMD Custom APU 0932, 16 GB | game folder on ext4, XDG_DATA_HOME an empty test folder on tmpfs; a test save in user/ with an unreadable entry after it |
first start in a copy of a test game folder, closed through the control bridge; https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/557 twice with the entry unreadable and once readable, main once unreadable | Test request T-WYKE: with https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/557 user/ unchanged both times and the new place empty, then the readable start moved everything; main lost 10 save files out of user/ into the new place while logging that the saves stay |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with #761’s change (perf/vk-occ-copy) applied; the renderer built from source (RX_VK_ONLY, release, llvm-mingw) into a nightly-80-7315957c game folder |
Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the defaults: Threaded rendering Auto (On: 20 logical processors), render 75% with Lanczos, frame generation off, V-Sync Auto; HOS_PROFILE=5, HOS_FPS_MAX=240 (not capped), RX_VK_FRAME_LOG |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 40 s after the teleport, then 40 s measured; one of 3 interleaved pairs (QS-1, QF-1, QF-2, QS-2, QS-3, QF-3); each from a fresh start; a test program holding logical processors 10 and 11 (the core kept for the render thread) at priority 15 for 35 ms of every 1000 ms in one burst, asleep between | run rt-qs-3 (5 of 6), the render thread kept to its core alone (before PR #838): average 55.1 fps, 1% low 18.8 fps, frame time median 16.78 ms and p99 50.85 ms over 2220 frames; 40 frames over 40 ms, the longest 57.1 ms; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with #761’s change (perf/vk-occ-copy) applied; the renderer built from source (RX_VK_ONLY, release, llvm-mingw) into a nightly-80-7315957c game folder |
Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the defaults: Threaded rendering Auto (On: 20 logical processors), render 75% with Lanczos, frame generation off, V-Sync Auto; HOS_PROFILE=5, HOS_FPS_MAX=240 (not capped), RX_VK_FRAME_LOG |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 40 s after the teleport, then 40 s measured; one of 3 interleaved pairs (QS-1, QF-1, QF-2, QS-2, QS-3, QF-3); each from a fresh start; a test program holding logical processors 10 and 11 (the core kept for the render thread) at priority 15 for 35 ms of every 1000 ms in one burst, asleep between | run rt-qs-2 (4 of 6), the render thread kept to its core alone (before PR #838): average 55.2 fps, 1% low 18.7 fps, frame time median 16.76 ms and p99 50.77 ms over 2220 frames; 40 frames over 40 ms, the longest 58.0 ms; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with #761’s change (perf/vk-occ-copy) applied; the renderer built from source (RX_VK_ONLY, release, llvm-mingw) into a nightly-80-7315957c game folder |
Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the defaults: Threaded rendering Auto (On: 20 logical processors), render 75% with Lanczos, frame generation off, V-Sync Auto; HOS_PROFILE=5, HOS_FPS_MAX=240 (not capped), RX_VK_FRAME_LOG |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 40 s after the teleport, then 40 s measured; one of 3 interleaved pairs (QS-1, QF-1, QF-2, QS-2, QS-3, QF-3); each from a fresh start; a test program holding logical processors 10 and 11 (the core kept for the render thread) at priority 15 for 35 ms of every 1000 ms in one burst, asleep between | run rt-qs-1 (1 of 6), the render thread kept to its core alone (before PR #838): average 55.3 fps, 1% low 19.0 fps, frame time median 16.76 ms and p99 51.05 ms over 2220 frames; 40 frames over 40 ms, the longest 56.1 ms; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with #761’s change (perf/vk-occ-copy) applied and the diff of fix/replay-thread-free-to-move @ 59389332 (PR #838) applied; the renderer built from source (RX_VK_ONLY, release, llvm-mingw) into a nightly-80-7315957c game folder |
Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the defaults: Threaded rendering Auto (On: 20 logical processors), render 75% with Lanczos, frame generation off, V-Sync Auto; HOS_PROFILE=5, HOS_FPS_MAX=240 (not capped), RX_VK_FRAME_LOG |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 40 s after the teleport, then 40 s measured; one of 3 interleaved pairs (QS-1, QF-1, QF-2, QS-2, QS-3, QF-3); each from a fresh start; a test program holding logical processors 10 and 11 (the core kept for the render thread) at priority 15 for 35 ms of every 1000 ms in one burst, asleep between | run rt-qf-3 (6 of 6), the render thread free to move (PR #838): average 57.4 fps, 1% low 43.2 fps, frame time median 16.71 ms and p99 22.41 ms over 2340 frames; 0 frames over 40 ms, the longest 26.9 ms; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with #761’s change (perf/vk-occ-copy) applied and the diff of fix/replay-thread-free-to-move @ 59389332 (PR #838) applied; the renderer built from source (RX_VK_ONLY, release, llvm-mingw) into a nightly-80-7315957c game folder |
Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the defaults: Threaded rendering Auto (On: 20 logical processors), render 75% with Lanczos, frame generation off, V-Sync Auto; HOS_PROFILE=5, HOS_FPS_MAX=240 (not capped), RX_VK_FRAME_LOG |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 40 s after the teleport, then 40 s measured; one of 3 interleaved pairs (QS-1, QF-1, QF-2, QS-2, QS-3, QF-3); each from a fresh start; a test program holding logical processors 10 and 11 (the core kept for the render thread) at priority 15 for 35 ms of every 1000 ms in one burst, asleep between | run rt-qf-2 (3 of 6), the render thread free to move (PR #838): average 57.4 fps, 1% low 41.9 fps, frame time median 16.74 ms and p99 22.20 ms over 2340 frames; 0 frames over 40 ms, the longest 28.2 ms; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with #761’s change (perf/vk-occ-copy) applied and the diff of fix/replay-thread-free-to-move @ 59389332 (PR #838) applied; the renderer built from source (RX_VK_ONLY, release, llvm-mingw) into a nightly-80-7315957c game folder |
Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the defaults: Threaded rendering Auto (On: 20 logical processors), render 75% with Lanczos, frame generation off, V-Sync Auto; HOS_PROFILE=5, HOS_FPS_MAX=240 (not capped), RX_VK_FRAME_LOG |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 40 s after the teleport, then 40 s measured; one of 3 interleaved pairs (QS-1, QF-1, QF-2, QS-2, QS-3, QF-3); each from a fresh start; a test program holding logical processors 10 and 11 (the core kept for the render thread) at priority 15 for 35 ms of every 1000 ms in one burst, asleep between | run rt-qf-1 (2 of 6), the render thread free to move (PR #838): average 57.2 fps, 1% low 41.2 fps, frame time median 16.78 ms and p99 22.76 ms over 2280 frames; 0 frames over 40 ms, the longest 30.4 ms; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with #761’s change (perf/vk-occ-copy) applied; the renderer built from source (RX_VK_ONLY, release, llvm-mingw) into a nightly-80-7315957c game folder |
Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the defaults: Threaded rendering Auto (On: 20 logical processors), render 75% with Lanczos, frame generation off, V-Sync Auto; HOS_PROFILE=5, HOS_FPS_MAX=240 (not capped), RX_VK_FRAME_LOG |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 40 s after the teleport, then 40 s measured; one of 3 interleaved pairs (PS-1, PF-1, PF-2, PS-2, PS-3, PF-3); each from a fresh start; Intel Graphics Software installed, nothing else started | run rt-ps-3 (5 of 6), the render thread kept to its core alone (before PR #838): average 57.8 fps, 1% low 42.3 fps, frame time median 16.61 ms and p99 22.52 ms over 2340 frames; 0 frames over 40 ms, the longest 26.1 ms; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with #761’s change (perf/vk-occ-copy) applied; the renderer built from source (RX_VK_ONLY, release, llvm-mingw) into a nightly-80-7315957c game folder |
Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the defaults: Threaded rendering Auto (On: 20 logical processors), render 75% with Lanczos, frame generation off, V-Sync Auto; HOS_PROFILE=5, HOS_FPS_MAX=240 (not capped), RX_VK_FRAME_LOG |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 40 s after the teleport, then 40 s measured; one of 3 interleaved pairs (PS-1, PF-1, PF-2, PS-2, PS-3, PF-3); each from a fresh start; Intel Graphics Software installed, nothing else started | run rt-ps-2 (4 of 6), the render thread kept to its core alone (before PR #838): average 57.5 fps, 1% low 38.5 fps, frame time median 16.67 ms and p99 22.74 ms over 2340 frames; 1 frames over 40 ms, the longest 52.6 ms; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with #761’s change (perf/vk-occ-copy) applied; the renderer built from source (RX_VK_ONLY, release, llvm-mingw) into a nightly-80-7315957c game folder |
Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the defaults: Threaded rendering Auto (On: 20 logical processors), render 75% with Lanczos, frame generation off, V-Sync Auto; HOS_PROFILE=5, HOS_FPS_MAX=240 (not capped), RX_VK_FRAME_LOG |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 40 s after the teleport, then 40 s measured; one of 3 interleaved pairs (PS-1, PF-1, PF-2, PS-2, PS-3, PF-3); each from a fresh start; Intel Graphics Software installed, nothing else started | run rt-ps-1 (1 of 6), the render thread kept to its core alone (before PR #838): average 57.3 fps, 1% low 41.1 fps, frame time median 16.71 ms and p99 22.89 ms over 2340 frames; 0 frames over 40 ms, the longest 27.7 ms; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with #761’s change (perf/vk-occ-copy) applied and the diff of fix/replay-thread-free-to-move @ 59389332 (PR #838) applied; the renderer built from source (RX_VK_ONLY, release, llvm-mingw) into a nightly-80-7315957c game folder |
Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the defaults: Threaded rendering Auto (On: 20 logical processors), render 75% with Lanczos, frame generation off, V-Sync Auto; HOS_PROFILE=5, HOS_FPS_MAX=240 (not capped), RX_VK_FRAME_LOG |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 40 s after the teleport, then 40 s measured; one of 3 interleaved pairs (PS-1, PF-1, PF-2, PS-2, PS-3, PF-3); each from a fresh start; Intel Graphics Software installed, nothing else started | run rt-pf-3 (6 of 6), the render thread free to move (PR #838): average 57.0 fps, 1% low 42.0 fps, frame time median 16.85 ms and p99 22.72 ms over 2280 frames; 0 frames over 40 ms, the longest 25.9 ms; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with #761’s change (perf/vk-occ-copy) applied and the diff of fix/replay-thread-free-to-move @ 59389332 (PR #838) applied; the renderer built from source (RX_VK_ONLY, release, llvm-mingw) into a nightly-80-7315957c game folder |
Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the defaults: Threaded rendering Auto (On: 20 logical processors), render 75% with Lanczos, frame generation off, V-Sync Auto; HOS_PROFILE=5, HOS_FPS_MAX=240 (not capped), RX_VK_FRAME_LOG |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 40 s after the teleport, then 40 s measured; one of 3 interleaved pairs (PS-1, PF-1, PF-2, PS-2, PS-3, PF-3); each from a fresh start; Intel Graphics Software installed, nothing else started | run rt-pf-2 (3 of 6), the render thread free to move (PR #838): average 57.0 fps, 1% low 42.8 fps, frame time median 16.89 ms and p99 22.57 ms over 2340 frames; 0 frames over 40 ms, the longest 26.5 ms; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with #761’s change (perf/vk-occ-copy) applied and the diff of fix/replay-thread-free-to-move @ 59389332 (PR #838) applied; the renderer built from source (RX_VK_ONLY, release, llvm-mingw) into a nightly-80-7315957c game folder |
Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the defaults: Threaded rendering Auto (On: 20 logical processors), render 75% with Lanczos, frame generation off, V-Sync Auto; HOS_PROFILE=5, HOS_FPS_MAX=240 (not capped), RX_VK_FRAME_LOG |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 40 s after the teleport, then 40 s measured; one of 3 interleaved pairs (PS-1, PF-1, PF-2, PS-2, PS-3, PF-3); each from a fresh start; Intel Graphics Software installed, nothing else started | run rt-pf-1 (2 of 6), the render thread free to move (PR #838): average 56.6 fps, 1% low 40.1 fps, frame time median 16.96 ms and p99 23.50 ms over 2280 frames; 0 frames over 40 ms, the longest 30.0 ms; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with #761’s change (perf/vk-occ-copy) applied; the renderer built from source (RX_VK_ONLY, release, llvm-mingw) into a nightly-80-7315957c game folder |
Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the defaults: Threaded rendering Auto (On: 20 logical processors), render 75% with Lanczos, frame generation off, V-Sync Auto; HOS_PROFILE=5, HOS_FPS_MAX=240 (not capped), RX_VK_FRAME_LOG |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 40 s after the teleport, then 40 s measured; one pair (HS-1, HF-1); each from a fresh start; a test program holding logical processors 10 and 11 (the core kept for the render thread) at priority 15, busy 35 ms of every 1000 ms and waking every 1 ms between | run rt-hs-1 (1 of 2), the render thread kept to its core alone (before PR #838): average 54.3 fps, 1% low 18.7 fps, frame time median 17.06 ms and p99 51.19 ms over 2160 frames; 39 frames over 40 ms, the longest 57.3 ms; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with #761’s change (perf/vk-occ-copy) applied and the diff of fix/replay-thread-free-to-move @ 59389332 (PR #838) applied; the renderer built from source (RX_VK_ONLY, release, llvm-mingw) into a nightly-80-7315957c game folder |
Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the defaults: Threaded rendering Auto (On: 20 logical processors), render 75% with Lanczos, frame generation off, V-Sync Auto; HOS_PROFILE=5, HOS_FPS_MAX=240 (not capped), RX_VK_FRAME_LOG |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 40 s after the teleport, then 40 s measured; one pair (HS-1, HF-1); each from a fresh start; a test program holding logical processors 10 and 11 (the core kept for the render thread) at priority 15, busy 35 ms of every 1000 ms and waking every 1 ms between | run rt-hf-1 (2 of 2), the render thread free to move (PR #838): average 44.6 fps, 1% low 11.9 fps, frame time median 21.66 ms and p99 34.24 ms over 1800 frames; 7 frames over 40 ms, the longest 500.3 ms; no device lost |
| 2026-11-15 | main a00c485c (dev_packs.py base a00c485c, no pack) |
CachyOS (Linux 7.3.3), headless sway (Wayland), fullscreen, 1280x720 output (the game’s choice) | RTX 3090 24 GB, NVIDIA 620.70.06 | Ryzen 9 7900X (the game on 12 of 24 threads: CPUs 0-5, 12-17), 32 GB | defaults (Threaded rendering Auto: on, the game’s own upscaler, no frame generation), HOS_FPS_MAX=240 (uncapped), RX_VK_FRAME_LOG on |
a fresh copy of one save in the starting forest each session: 15 s standing, 20 s turning the camera (right stick held), then the debug teleport to the main outdoor area and to the Market, each after 15 s: 15 s standing, 20 s turning; frame times from RX_VK_FRAME_LOG; six sessions interleaved C M M C C M |
Session 6 of 6 (10:41 UTC, main): the starting forest standing 12.28 ms median, p99 13.8 ms, 81.0 fps; the starting forest turning 11.31 ms median, p99 15.9 ms, 86.6 fps; the main outdoor area standing 8.72 ms median, p99 10.5 ms, 113.7 fps; the main outdoor area turning 8.85 ms median, p99 11.4 ms, 112.1 fps; Market standing 12.49 ms median, p99 19.5 ms, 77.9 fps; Market turning 11.33 ms median, p99 18.6 ms, 84.0 fps; no crash, no device lost |
| 2026-11-15 | main a00c485c with this PR’s host/nvn_native.c change (dev_packs.py pack copylock on base a00c485c) |
CachyOS (Linux 7.3.3), headless sway (Wayland), fullscreen, 1280x720 output (the game’s choice) | RTX 3090 24 GB, NVIDIA 620.70.06 | Ryzen 9 7900X (the game on 12 of 24 threads: CPUs 0-5, 12-17), 32 GB | defaults (Threaded rendering Auto: on, the game’s own upscaler, no frame generation), HOS_FPS_MAX=240 (uncapped), RX_VK_FRAME_LOG on |
a fresh copy of one save in the starting forest each session: 15 s standing, 20 s turning the camera (right stick held), then the debug teleport to the main outdoor area and to the Market, each after 15 s: 15 s standing, 20 s turning; frame times from RX_VK_FRAME_LOG; six sessions interleaved C M M C C M |
Session 5 of 6 (10:37 UTC, this PR): the starting forest standing 12.20 ms median, p99 13.9 ms, 81.4 fps; the starting forest turning 11.20 ms median, p99 14.3 ms, 89.0 fps; the main outdoor area standing 8.68 ms median, p99 11.5 ms, 114.3 fps; the main outdoor area turning 8.72 ms median, p99 12.2 ms, 114.5 fps; Market standing 12.30 ms median, p99 19.7 ms, 77.8 fps; Market turning 10.85 ms median, p99 17.5 ms, 94.3 fps; no crash, no device lost |
| 2026-11-15 | main a00c485c with this PR’s host/nvn_native.c change (dev_packs.py pack copylock on base a00c485c) |
CachyOS (Linux 7.3.3), headless sway (Wayland), fullscreen, 1280x720 output (the game’s choice) | RTX 3090 24 GB, NVIDIA 620.70.06 | Ryzen 9 7900X (the game on 12 of 24 threads: CPUs 0-5, 12-17), 32 GB | defaults (Threaded rendering Auto: on, the game’s own upscaler, no frame generation), HOS_FPS_MAX=240 (uncapped), RX_VK_FRAME_LOG on |
a fresh copy of one save in the starting forest each session: 15 s standing, 20 s turning the camera (right stick held), then the debug teleport to the main outdoor area and to the Market, each after 15 s: 15 s standing, 20 s turning; frame times from RX_VK_FRAME_LOG; six sessions interleaved C M M C C M |
Session 4 of 6 (10:33 UTC, this PR): the starting forest standing 12.22 ms median, p99 13.6 ms, 81.5 fps; the starting forest turning 11.21 ms median, p99 14.8 ms, 89.0 fps; the main outdoor area standing 8.60 ms median, p99 9.9 ms, 115.3 fps; the main outdoor area turning 8.88 ms median, p99 12.2 ms, 113.5 fps; Market standing 12.09 ms median, p99 17.7 ms, 79.9 fps; Market turning 11.11 ms median, p99 18.7 ms, 90.8 fps; no crash, no device lost |
| 2026-11-15 | main a00c485c (dev_packs.py base a00c485c, no pack) |
CachyOS (Linux 7.3.3), headless sway (Wayland), fullscreen, 1280x720 output (the game’s choice) | RTX 3090 24 GB, NVIDIA 620.70.06 | Ryzen 9 7900X (the game on 12 of 24 threads: CPUs 0-5, 12-17), 32 GB | defaults (Threaded rendering Auto: on, the game’s own upscaler, no frame generation), HOS_FPS_MAX=240 (uncapped), RX_VK_FRAME_LOG on |
a fresh copy of one save in the starting forest each session: 15 s standing, 20 s turning the camera (right stick held), then the debug teleport to the main outdoor area and to the Market, each after 15 s: 15 s standing, 20 s turning; frame times from RX_VK_FRAME_LOG; six sessions interleaved C M M C C M |
Session 3 of 6 (10:30 UTC, main): the starting forest standing 12.34 ms median, p99 14.8 ms, 80.3 fps; the starting forest turning 11.36 ms median, p99 14.7 ms, 87.9 fps; the main outdoor area standing 8.70 ms median, p99 10.7 ms, 114.0 fps; the main outdoor area turning 8.88 ms median, p99 12.4 ms, 111.5 fps; Market standing 12.34 ms median, p99 21.1 ms, 78.0 fps; Market turning 11.07 ms median, p99 18.1 ms, 91.1 fps; no crash, no device lost |
| 2026-11-15 | main a00c485c (dev_packs.py base a00c485c, no pack) |
CachyOS (Linux 7.3.3), headless sway (Wayland), fullscreen, 1280x720 output (the game’s choice) | RTX 3090 24 GB, NVIDIA 620.70.06 | Ryzen 9 7900X (the game on 12 of 24 threads: CPUs 0-5, 12-17), 32 GB | defaults (Threaded rendering Auto: on, the game’s own upscaler, no frame generation), HOS_FPS_MAX=240 (uncapped), RX_VK_FRAME_LOG on |
a fresh copy of one save in the starting forest each session: 15 s standing, 20 s turning the camera (right stick held), then the debug teleport to the main outdoor area and to the Market, each after 15 s: 15 s standing, 20 s turning; frame times from RX_VK_FRAME_LOG; six sessions interleaved C M M C C M |
Session 2 of 6 (10:26 UTC, main): the starting forest standing 12.34 ms median, p99 13.7 ms, 80.6 fps; the starting forest turning 11.42 ms median, p99 16.5 ms, 86.2 fps; the main outdoor area standing 8.65 ms median, p99 10.7 ms, 114.7 fps; the main outdoor area turning 8.99 ms median, p99 12.6 ms, 109.4 fps; Market standing 12.71 ms median, p99 17.4 ms, 77.3 fps; Market turning 12.25 ms median, p99 19.9 ms, 76.2 fps; no crash, no device lost |
| 2026-11-15 | main a00c485c with this PR’s host/nvn_native.c change (dev_packs.py pack copylock on base a00c485c) |
CachyOS (Linux 7.3.3), headless sway (Wayland), fullscreen, 1280x720 output (the game’s choice) | RTX 3090 24 GB, NVIDIA 620.70.06 | Ryzen 9 7900X (the game on 12 of 24 threads: CPUs 0-5, 12-17), 32 GB | defaults (Threaded rendering Auto: on, the game’s own upscaler, no frame generation), HOS_FPS_MAX=240 (uncapped), RX_VK_FRAME_LOG on |
a fresh copy of one save in the starting forest each session: 15 s standing, 20 s turning the camera (right stick held), then the debug teleport to the main outdoor area and to the Market, each after 15 s: 15 s standing, 20 s turning; frame times from RX_VK_FRAME_LOG; six sessions interleaved C M M C C M |
Session 1 of 6 (10:22 UTC, this PR): the starting forest standing 12.38 ms median, p99 14.7 ms, 80.5 fps; the starting forest turning 11.30 ms median, p99 16.1 ms, 86.8 fps; the main outdoor area standing 8.62 ms median, p99 10.8 ms, 115.1 fps; the main outdoor area turning 8.66 ms median, p99 11.5 ms, 115.7 fps; Market standing 12.31 ms median, p99 20.0 ms, 77.9 fps; Market turning 10.89 ms median, p99 18.0 ms, 92.2 fps; no crash, no device lost |
| 2026-11-15 | main 7ff5256e against the combined build of main 7ff5256e with https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/550 to https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/558 merged (https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/553 @ 93c285b7, https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/554 @ 6c76d0eb), x86-64-v3, release play profile, LLVM 23.1.5 | SteamOS 3.9.5 (build 20261023.100), kernel 7.2.7-valve2, Game Mode, docked | AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel git-5c65d4fc31 | Steam Deck OLED, AMD Custom APU 0932, 16 GB | Minimum (4GB) preset: render 50% of 1280x720, 2.2 GB VRAM budget, Handheld profile, the game’s own upscaler, frame generation off, stock clocks | the starting forest and 10 debug teleports through the control bridge, about 45 s an area, about 11 minutes a run; four runs interleaved combined, main, main, combined; fps from the bridge every 5 s | Test requests T-4H95 and T-GQD8 (one run set): fps 16.8 / 17.1 with the fixes, 16.6 / 17.7 on main; no did not complete in time line (the timed-out read-back never happened: a non-regression run); no device lost, fault, stall, crash or kernel GPU reset; 14 to 17 textures evicted and made again a run; no garbled textures in 44 arrival pictures |
| 2026-11-15 | local integration vrs-verify bfc69759: main aa95d35f with this PR (precompile-save-before-title), radv-vrs-key-switch and driver-cache-diagnostics, a game pack built via scripts/dev_packs.py on a base of main 9c9d69ed; the same pack for both starts |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2), headless | Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), stock GPU clock | Steam Deck OLED APU, 16 GB RAM | Defaults, the precompile at its default and RX_VK_PRECOMPILE_SAVE_S unset (30); RADV_FORCE_VRS_CONFIG_FILE naming a file holding 1x1 and RX_VK_HALF_RATE_SHADING=1, so the variable is kept and RADV keys every graphics pipeline differently from the caches; a scratch game folder starting from a fresh copy of warm caches made without the variable (32,438 noted graphics pipelines) |
The start-up to the title, two starts in a row from the same folder (run deck02): c1 from the fresh copy, c2 from what c1 left | c1: every pipeline compiled again at about 20 a second (31,570 of 38,345 noted by 1,342 s), precompile 1,466,663 ms, then vk: pipeline cache saved (449813 KB) and vk: precompile: its pipelines took 1454 s (compiled, not found in the caches): the pipeline cache written, 439.3 MB, in 821 ms, before the title (RX_VK_PRECOMPILE_SAVE_S=0: not); title at 1,501 s; the cache file 456,050,920 to 460,608,696 bytes. c2: nothing compiled, precompile 13,355 ms, title at 46 s. Without this PR (main aa95d35f with other open PRs, a start that compiled for 1,222 s) the cache grew 4.3 MB and was not written (written once it is 64 MB larger). The start-up screen’s saving state was not seen (headless) |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with the diff of perf/vk-occ-copy @ c0da900e applied, the renderer built from source into a nightly-80-7315957c game folder | Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the default: the copies on (Intel’s Windows driver, a discrete GPU); HOS_PROFILE=5, RX_VK_OCC_LOG=1, RX_VK_FRAME_LOG; render 75% with Lanczos, frame generation off, V-Sync Auto, the console’s 60 fps cap (not reached) |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 45 s after the teleport, then 45 s measured; one of 3 interleaved pairs (on, off, off, on, on, off), each from a fresh start | run P1-3 (5 of 6): average 46.1 fps, 1% low 27.9 fps, frame time median 20.59 ms and p99 34.22 ms over 2100 frames; GPU time 12.2 ms a frame; the counts’ reads 0.09 ms a frame; 12.4 copies a frame; 53.5% of the reports read 0; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with the diff of perf/vk-occ-copy @ c0da900e applied, the renderer built from source into a nightly-80-7315957c game folder | Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the default: the copies on (Intel’s Windows driver, a discrete GPU); HOS_PROFILE=5, RX_VK_OCC_LOG=1, RX_VK_FRAME_LOG; render 75% with Lanczos, frame generation off, V-Sync Auto, the console’s 60 fps cap (not reached) |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 45 s after the teleport, then 45 s measured; one of 3 interleaved pairs (on, off, off, on, on, off), each from a fresh start | run P1-2 (4 of 6): average 45.6 fps, 1% low 25.3 fps, frame time median 20.41 ms and p99 37.71 ms over 2040 frames; GPU time 12.2 ms a frame; the counts’ reads 0.11 ms a frame; 14.0 copies a frame; 53.5% of the reports read 0; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with the diff of perf/vk-occ-copy @ c0da900e applied, the renderer built from source into a nightly-80-7315957c game folder | Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the default: the copies on (Intel’s Windows driver, a discrete GPU); HOS_PROFILE=5, RX_VK_OCC_LOG=1, RX_VK_FRAME_LOG; render 75% with Lanczos, frame generation off, V-Sync Auto, the console’s 60 fps cap (not reached) |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 45 s after the teleport, then 45 s measured; one of 3 interleaved pairs (on, off, off, on, on, off), each from a fresh start | run P1-1 (1 of 6): average 45.8 fps, 1% low 23.9 fps, frame time median 20.21 ms and p99 39.37 ms over 2100 frames; GPU time 12.1 ms a frame; the counts’ reads 0.11 ms a frame; 14.0 copies a frame; 53.7% of the reports read 0; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with the diff of perf/vk-occ-copy @ c0da900e applied, the renderer built from source into a nightly-80-7315957c game folder | Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | RX_VK_OCC_COPY=0: the query pool read, as on main; HOS_PROFILE=5, RX_VK_OCC_LOG=1, RX_VK_FRAME_LOG; render 75% with Lanczos, frame generation off, V-Sync Auto, the console’s 60 fps cap (not reached) |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 45 s after the teleport, then 45 s measured; one of 3 interleaved pairs (on, off, off, on, on, off), each from a fresh start | run P0-3 (6 of 6): average 34.8 fps, 1% low 24.1 fps, frame time median 27.59 ms and p99 40.75 ms over 1560 frames; GPU time 10.2 ms a frame; the counts’ reads 7.43 ms a frame; 0.0 copies a frame; 53.7% of the reports read 0; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with the diff of perf/vk-occ-copy @ c0da900e applied, the renderer built from source into a nightly-80-7315957c game folder | Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | RX_VK_OCC_COPY=0: the query pool read, as on main; HOS_PROFILE=5, RX_VK_OCC_LOG=1, RX_VK_FRAME_LOG; render 75% with Lanczos, frame generation off, V-Sync Auto, the console’s 60 fps cap (not reached) |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 45 s after the teleport, then 45 s measured; one of 3 interleaved pairs (on, off, off, on, on, off), each from a fresh start | run P0-2 (3 of 6): average 35.5 fps, 1% low 22.7 fps, frame time median 26.95 ms and p99 40.70 ms over 1620 frames; GPU time 10.2 ms a frame; the counts’ reads 7.42 ms a frame; 0.0 copies a frame; 53.9% of the reports read 0; no device lost |
| 2026-11-15 | 690d8446 (main 2b1d38c6 with #443, #446, #459 and #500 merged) with the diff of perf/vk-occ-copy @ c0da900e applied, the renderer built from source into a nightly-80-7315957c game folder | Windows 11 Pro 10.0.29704, 1280x720 window | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | RX_VK_OCC_COPY=0: the query pool read, as on main; HOS_PROFILE=5, RX_VK_OCC_LOG=1, RX_VK_FRAME_LOG; render 75% with Lanczos, frame generation off, V-Sync Auto, the console’s 60 fps cap (not reached) |
Castle Town’s market as child by day, standing at the entrance (the control bridge’s teleport Market_Child); 45 s after the teleport, then 45 s measured; one of 3 interleaved pairs (on, off, off, on, on, off), each from a fresh start | run P0-1 (2 of 6): average 33.7 fps, 1% low 21.6 fps, frame time median 28.23 ms and p99 43.26 ms over 1560 frames; GPU time 10.2 ms a frame; the counts’ reads 7.56 ms a frame; 0.0 copies a frame; 54.1% of the reports read 0; no device lost |
| 2026-11-15 UTC | Renderer a3e58d2f: the Most-to-All scheduling fix backported over f7fc0b73 on base 82886e89; base host unchanged | Windows 11 Home N 26300, preparation only | AMD Radeon RX 9070 XT 16 GB, AMD proprietary 26.10.2 (LLPC) | Ryzen 7 5700X3D, 64 GB | --prepare-shaders --compile-shaders all, RX_VK_PREDICT_S=1; base renderer PGO preserved; real prepared cache, prediction-completion files withheld in the test runtime to match the Most-prepared base; no cache header edited |
Before the title, native preparation CLI; tester klimovich008, run02 | Exit 0; prediction entered and enumerated 60,399 candidates. The one-second limit expired while deciding them, so zero predicted pipelines were built. Noted and compute warm-up completed with zero not made. This proves the scheduling change only; no gameplay, speed or complete-preparation claim. |
| 2026-11-15 UTC | Renderer f7fc0b73 over base 82886e89; the pipeline-cache and compute-progress changes are present in both comparison arms | Windows 11 Home N 26300, preparation only | AMD Radeon RX 9070 XT 16 GB, AMD proprietary 26.10.2 (LLPC) | Ryzen 7 5700X3D, 64 GB | --prepare-shaders --compile-shaders all, RX_VK_PREDICT_S=1; renderer PGO absent; real prepared cache, prediction-completion files withheld in the test runtime to match the Most-prepared base; no cache header edited |
Before the title, native preparation CLI; tester klimovich008, run01 | Exit 0 without entering prediction; prediction-completion keys still absent. Noted and compute warm-up completed with zero not made. This checks scheduling, not gameplay or speed; no full-prediction claim. |
| 2026-11-15 | ed097a30aeaa424cc25a7d9ade43b110312877fc | Windows; OS build unknown; no window | Not used; game-free runner | CPU/RAM not recorded | Python 3.14; LLVM-MinGW on PATH; RX_CPM set; CXX clang++; 4 jobs; default platform skips | tests/run_all.py, full rebased-topic run | win-suite01: 537 passed, 21 failed, 35 skipped in 287 s; 20 failures also occur on main; remaining control-tool failure passed on an isolated topic rerun; not an all-green suite |
| 2026-11-15 | ed097a30aeaa424cc25a7d9ade43b110312877fc | Windows; OS build unknown; no window | Not used; game-free fixture | CPU/RAM not recorded | Python 3.14; same tool environment as win-suite01; one selected check | tests/run_all.py restricted to tests/test_longshot_ctl.py | win-retry01: control-tool check passed in 48 s; original full-run failure read an empty completion receipt; intermittent cause not established |
| 2026-11-15 | main 9e6f231d0ded199793e8270d20ec60cdb765d28f production sources plus the new ed097a30 test | Windows; OS build unknown; no window | Not used; game-free fixture | CPU/RAM not recorded | Python 3.14; LLVM-MinGW 20261104; fake SDL; one warm-up per variant excluded | tests/test_audio_in.py, measured sequence main/fix/main/fix, invocation 3 | win-pair02-main: expected expanded-list/boundary failure; native exit 43, Python exit 1; 0.398 s diagnostic elapsed; no performance claim |
| 2026-11-15 | ed097a30aeaa424cc25a7d9ade43b110312877fc; documentation-only working changes | Windows; OS build unknown; no window | Not used; game-free fixture | CPU/RAM not recorded | Python 3.14; LLVM-MinGW 20261104; fake SDL; one warm-up per variant excluded | tests/test_audio_in.py, measured sequence main/fix/main/fix, invocation 4 | win-pair02-fix: Python exit 0; 55 host cases and expanded-device menu fixture pass; 0.896 s diagnostic elapsed; no real mic capture or performance claim |
| 2026-11-15 | main 9e6f231d0ded199793e8270d20ec60cdb765d28f production sources plus the new ed097a30 test | Windows; OS build unknown; no window | Not used; game-free fixture | CPU/RAM not recorded | Python 3.14; LLVM-MinGW 20261104; fake SDL; one warm-up per variant excluded | tests/test_audio_in.py, measured sequence main/fix/main/fix, invocation 1 | win-pair01-main: expected expanded-list/boundary failure; native exit 43, Python exit 1; 0.398 s diagnostic elapsed; no performance claim |
| 2026-11-15 | ed097a30aeaa424cc25a7d9ade43b110312877fc; documentation-only working changes | Windows; OS build unknown; no window | Not used; game-free fixture | CPU/RAM not recorded | Python 3.14; LLVM-MinGW 20261104; fake SDL; one warm-up per variant excluded | tests/test_audio_in.py, measured sequence main/fix/main/fix, invocation 2 | win-pair01-fix: Python exit 0; 55 host cases and expanded-device menu fixture pass; 0.969 s diagnostic elapsed; no real mic capture or performance claim |
| 2026-11-15 | ed097a30aeaa424cc25a7d9ade43b110312877fc | Windows; OS build unknown; no window | Not used; game-free fixture | CPU/RAM not recorded | Python 3.14; LLVM-MinGW 20261104; fake SDL; 70 recording endpoints | tests/test_audio_in.py, rebased mic-limit branch | win-fix01: 55 native input-stream cases and expanded-device menu fixture passed; includes high index, endpoint 63, truncation, reconnect and Default; no real audio capture |
| 2026-11-15 | main 9e6f231d0ded199793e8270d20ec60cdb765d28f production sources with the new test from ed097a30 | Windows; OS build unknown; no window | Not used; game-free fixture | CPU/RAM not recorded | Python 3.14; LLVM-MinGW 20261104; fake SDL; 70 recording endpoints | tests/test_audio_in.py, unchanged-main control | win-control01: expected failure at expanded-list/boundary assertion, native exit 43; Python driver exit 1; no real audio capture |
| 2026-11-15 | main 9e6f231d0ded199793e8270d20ec60cdb765d28f | Windows; OS build unknown; no window | Not used; game-free runner | CPU/RAM not recorded | Python 3.14; LLVM-MinGW on PATH; RX_CPM set; CXX clang++; 4 jobs | tests/run_all.py restricted to the topic full-run failures | win-baseline01: 1 passed, 20 failed, 0 skipped in 52 s; the passing check was test_longshot_ctl.py; baseline failures are named in the PR description |
| 2026-11-15 | main 9e6f231d0ded199793e8270d20ec60cdb765d28f production sources plus the new ed097a30 test | Ubuntu 24.04.5 WSL2; kernel 6.18.58.1-microsoft-standard-WSL2; no window | Not used; game-free fixture | CPU/RAM not recorded | Python 3.12.3; GCC/G++ 13.3.0; fake SDL; one warm-up per variant excluded | tests/test_audio_in.py, measured sequence main/fix/main/fix, invocation 3 | linux-pair02-main: expected expanded-list/boundary failure; native exit 43, Python exit 1; 0.617 s diagnostic elapsed; no performance claim |
| 2026-11-15 | ed097a30aeaa424cc25a7d9ade43b110312877fc; documentation-only working changes | Ubuntu 24.04.5 WSL2; kernel 6.18.58.1-microsoft-standard-WSL2; no window | Not used; game-free fixture | CPU/RAM not recorded | Python 3.12.3; GCC/G++ 13.3.0; fake SDL; one warm-up per variant excluded | tests/test_audio_in.py, measured sequence main/fix/main/fix, invocation 4 | linux-pair02-fix: Python exit 0; 55 host cases and expanded-device menu fixture pass; 1.289 s diagnostic elapsed; no real mic capture or performance claim |
| 2026-11-15 | main 9e6f231d0ded199793e8270d20ec60cdb765d28f production sources plus the new ed097a30 test | Ubuntu 24.04.5 WSL2; kernel 6.18.58.1-microsoft-standard-WSL2; no window | Not used; game-free fixture | CPU/RAM not recorded | Python 3.12.3; GCC/G++ 13.3.0; fake SDL; one warm-up per variant excluded | tests/test_audio_in.py, measured sequence main/fix/main/fix, invocation 1 | linux-pair01-main: expected expanded-list/boundary failure; native exit 43, Python exit 1; 0.686 s diagnostic elapsed; no performance claim |
| 2026-11-15 | ed097a30aeaa424cc25a7d9ade43b110312877fc; documentation-only working changes | Ubuntu 24.04.5 WSL2; kernel 6.18.58.1-microsoft-standard-WSL2; no window | Not used; game-free fixture | CPU/RAM not recorded | Python 3.12.3; GCC/G++ 13.3.0; fake SDL; one warm-up per variant excluded | tests/test_audio_in.py, measured sequence main/fix/main/fix, invocation 2 | linux-pair01-fix: Python exit 0; 55 host cases and expanded-device menu fixture pass; 1.319 s diagnostic elapsed; no real mic capture or performance claim |
| 2026-11-15 | ed097a30aeaa424cc25a7d9ade43b110312877fc | Ubuntu 24.04.5 WSL2; kernel 6.18.58.1-microsoft-standard-WSL2; no window | Not used; game-free fixture | CPU/RAM not recorded | Python 3.12.3; GCC/G++ 13.3.0; fake SDL; 70 recording endpoints | tests/test_audio_in.py, rebased mic-limit branch | linux-fix01: 55 native input-stream cases and expanded-device menu fixture passed with warnings treated as errors; includes boundary/default/reconnect cases; no full host build or real audio capture |
| 2026-11-15 | main 9e6f231d0ded199793e8270d20ec60cdb765d28f production sources with the new test from ed097a30 | Ubuntu 24.04.5 WSL2; kernel 6.18.58.1-microsoft-standard-WSL2; no window | Not used; game-free fixture | CPU/RAM not recorded | Python 3.12.3; GCC/G++ 13.3.0; fake SDL; 70 recording endpoints | tests/test_audio_in.py, unchanged-main control | linux-control01: expected failure at expanded-list/boundary assertion, native exit 43; Python driver exit 1; no real audio capture |
| 2026-11-15 | longshot v3.0.0-rc.4-11-ga00c485c-dirty (source 0b226774): main a00c485c with this change’s two fixes in the working tree (not committed when run) and a local counter of dark frames (not in the change); built with make_game.py |
Windows 11 Pro 26200, no window (headless, scripted input) | RTX 4070 SUPER 12 GB, NVIDIA 617.42 | Ryzen 9 9950X3D (16 cores, SMT off), 32 GB | 2560x1440, DLAA, no frame generation, uncapped (HOS_FPS_MAX=1000); RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32, RX_VK_LOAD=create, RX_VK_TEST_ALLOWED_MB=10200; the render thread (RX_VK_ASYNC=1, what Threaded rendering’s Auto picks on this PC) |
Castle Town’s south gate, save set “First Step into the game world Town”: out to the drawbridge and back through the gate several times on a wall-clock script (a few loading screens), night at 90 s, 140 s; a fourth round, after the change was moved onto main a00c485c, taken in turn with the two other sides | run ta4 (round 4, the one on main a00c485c): 0 dark-Link frames (frames drawn with an empty light grid, counted in the build) in 28197 frames; 0 hold timeouts (50 ms in the log’s order line); mean 197.7 fps over the run, loads included; 54 frames over 50 ms; 20071 captures at a CPU event’s wait, 0 of them at a wait met on arrival; stopped queues held for a GPU wait 0 times |
| 2026-11-15 | longshot v3.0.0-rc.4-1-ge090cc48-dirty (source 09e03930): main e090cc48 with this change’s two fixes in the working tree (not committed when run) and a local counter of dark frames (not in the change); built with make_game.py |
Windows 11 Pro 26200, no window (headless, scripted input) | RTX 4070 SUPER 12 GB, NVIDIA 617.42 | Ryzen 9 9950X3D (16 cores, SMT off), 32 GB | 2560x1440, DLAA, no frame generation, uncapped (HOS_FPS_MAX=1000); RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32, RX_VK_LOAD=create, RX_VK_TEST_ALLOWED_MB=10200; the render thread (RX_VK_ASYNC=1, what Threaded rendering’s Auto picks on this PC) |
Castle Town’s south gate, save set “First Step into the game world Town”: out to the drawbridge and back through the gate several times on a wall-clock script (a few loading screens), night at 90 s, 140 s; one of three rounds taken in turn with the two other sides | run ta3 (round 3 of 3): 0 dark-Link frames (frames drawn with an empty light grid, counted in the build) in 22370 frames; 1 hold timeouts (50 ms in the log’s order line); mean 146.5 fps over the run, loads included; 29 frames over 50 ms; 16569 captures at a CPU event’s wait, 0 of them at a wait met on arrival; stopped queues held for a GPU wait 0 times |
| 2026-11-15 | longshot v3.0.0-rc.4-1-ge090cc48-dirty (source 09e03930): main e090cc48 with this change’s two fixes in the working tree (not committed when run) and a local counter of dark frames (not in the change); built with make_game.py |
Windows 11 Pro 26200, no window (headless, scripted input) | RTX 4070 SUPER 12 GB, NVIDIA 617.42 | Ryzen 9 9950X3D (16 cores, SMT off), 32 GB | 2560x1440, DLAA, no frame generation, uncapped (HOS_FPS_MAX=1000); RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32, RX_VK_LOAD=create, RX_VK_TEST_ALLOWED_MB=10200; the render thread (RX_VK_ASYNC=1, what Threaded rendering’s Auto picks on this PC) |
Castle Town’s south gate, save set “First Step into the game world Town”: out to the drawbridge and back through the gate several times on a wall-clock script (a few loading screens), night at 90 s, 140 s; one of three rounds taken in turn with the two other sides | run ta2 (round 2 of 3): 0 dark-Link frames (frames drawn with an empty light grid, counted in the build) in 21326 frames; 0 hold timeouts (50 ms in the log’s order line); mean 138.5 fps over the run, loads included; 34 frames over 50 ms; 16399 captures at a CPU event’s wait, 0 of them at a wait met on arrival; stopped queues held for a GPU wait 0 times |
| 2026-11-15 | longshot v3.0.0-rc.4-1-ge090cc48-dirty (source 09e03930): main e090cc48 with this change’s two fixes in the working tree (not committed when run) and a local counter of dark frames (not in the change); built with make_game.py |
Windows 11 Pro 26200, no window (headless, scripted input) | RTX 4070 SUPER 12 GB, NVIDIA 617.42 | Ryzen 9 9950X3D (16 cores, SMT off), 32 GB | 2560x1440, DLAA, no frame generation, uncapped (HOS_FPS_MAX=1000); RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32, RX_VK_LOAD=create, RX_VK_TEST_ALLOWED_MB=10200; the render thread (RX_VK_ASYNC=1, what Threaded rendering’s Auto picks on this PC) |
Castle Town’s south gate, save set “First Step into the game world Town”: out to the drawbridge and back through the gate several times on a wall-clock script (a few loading screens), night at 90 s, 140 s; one of three rounds taken in turn with the two other sides | run ta1 (round 1 of 3): 0 dark-Link frames (frames drawn with an empty light grid, counted in the build) in 22000 frames; 0 hold timeouts (50 ms in the log’s order line); mean 143.7 fps over the run, loads included; 33 frames over 50 ms; 16331 captures at a CPU event’s wait, 0 of them at a wait met on arrival; stopped queues held for a GPU wait 0 times |
| 2026-11-15 | longshot v3.0.0-rc.4-11-ga00c485c-dirty (source 0b226774): main a00c485c with this change’s two fixes in the working tree (not committed when run) and a local counter of dark frames (not in the change); built with make_game.py |
Windows 11 Pro 26200, no window (headless, scripted input) | RTX 4070 SUPER 12 GB, NVIDIA 617.42 | Ryzen 9 9950X3D (16 cores, SMT off), 32 GB | 2560x1440, DLAA, no frame generation, uncapped (HOS_FPS_MAX=1000); RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32, RX_VK_LOAD=create, RX_VK_TEST_ALLOWED_MB=10200; the submitting thread with both changes on, the default (RX_VK_ASYNC=0) |
Castle Town’s south gate, save set “First Step into the game world Town”: out to the drawbridge and back through the gate several times on a wall-clock script (a few loading screens), night at 90 s, 140 s; a fourth round, after the change was moved onto main a00c485c, taken in turn with the two other sides | run sb4 (round 4, the one on main a00c485c): 0 dark-Link frames (frames drawn with an empty light grid, counted in the build) in 19849 frames; 0 hold timeouts (50 ms in the log’s order line); mean 130.5 fps over the run, loads included; 36 frames over 50 ms; 6125 captures at a CPU event’s wait, 408 of them at a wait met on arrival; stopped queues held for a GPU wait 2 times |
| 2026-11-15 | longshot v3.0.0-rc.4-1-ge090cc48-dirty (source 09e03930): main e090cc48 with this change’s two fixes in the working tree (not committed when run) and a local counter of dark frames (not in the change); built with make_game.py |
Windows 11 Pro 26200, no window (headless, scripted input) | RTX 4070 SUPER 12 GB, NVIDIA 617.42 | Ryzen 9 9950X3D (16 cores, SMT off), 32 GB | 2560x1440, DLAA, no frame generation, uncapped (HOS_FPS_MAX=1000); RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32, RX_VK_LOAD=create, RX_VK_TEST_ALLOWED_MB=10200; the submitting thread with both changes on, the default (RX_VK_ASYNC=0) |
Castle Town’s south gate, save set “First Step into the game world Town”: out to the drawbridge and back through the gate several times on a wall-clock script (a few loading screens), night at 90 s, 140 s; one of three rounds taken in turn with the two other sides | run sb3 (round 3 of 3): 0 dark-Link frames (frames drawn with an empty light grid, counted in the build) in 20210 frames; 0 hold timeouts (50 ms in the log’s order line); mean 129.5 fps over the run, loads included; 40 frames over 50 ms; 6594 captures at a CPU event’s wait, 478 of them at a wait met on arrival; stopped queues held for a GPU wait 1 times |
| 2026-11-15 | longshot v3.0.0-rc.4-1-ge090cc48-dirty (source 09e03930): main e090cc48 with this change’s two fixes in the working tree (not committed when run) and a local counter of dark frames (not in the change); built with make_game.py |
Windows 11 Pro 26200, no window (headless, scripted input) | RTX 4070 SUPER 12 GB, NVIDIA 617.42 | Ryzen 9 9950X3D (16 cores, SMT off), 32 GB | 2560x1440, DLAA, no frame generation, uncapped (HOS_FPS_MAX=1000); RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32, RX_VK_LOAD=create, RX_VK_TEST_ALLOWED_MB=10200; the submitting thread with both changes on, the default (RX_VK_ASYNC=0) |
Castle Town’s south gate, save set “First Step into the game world Town”: out to the drawbridge and back through the gate several times on a wall-clock script (a few loading screens), night at 90 s, 140 s; one of three rounds taken in turn with the two other sides | run sb2 (round 2 of 3): 0 dark-Link frames (frames drawn with an empty light grid, counted in the build) in 20493 frames; 0 hold timeouts (50 ms in the log’s order line); mean 132.2 fps over the run, loads included; 37 frames over 50 ms; 6288 captures at a CPU event’s wait, 743 of them at a wait met on arrival; stopped queues held for a GPU wait 2 times |
| 2026-11-15 | longshot v3.0.0-rc.4-1-ge090cc48-dirty (source 09e03930): main e090cc48 with this change’s two fixes in the working tree (not committed when run) and a local counter of dark frames (not in the change); built with make_game.py |
Windows 11 Pro 26200, no window (headless, scripted input) | RTX 4070 SUPER 12 GB, NVIDIA 617.42 | Ryzen 9 9950X3D (16 cores, SMT off), 32 GB | 2560x1440, DLAA, no frame generation, uncapped (HOS_FPS_MAX=1000); RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32, RX_VK_LOAD=create, RX_VK_TEST_ALLOWED_MB=10200; the submitting thread with both changes on, the default (RX_VK_ASYNC=0) |
Castle Town’s south gate, save set “First Step into the game world Town”: out to the drawbridge and back through the gate several times on a wall-clock script (a few loading screens), night at 90 s, 140 s; one of three rounds taken in turn with the two other sides | run sb1 (round 1 of 3): 0 dark-Link frames (frames drawn with an empty light grid, counted in the build) in 18795 frames; 0 hold timeouts (50 ms in the log’s order line); mean 118.5 fps over the run, loads included; 35 frames over 50 ms; 6710 captures at a CPU event’s wait, 526 of them at a wait met on arrival; stopped queues held for a GPU wait 1 times |
| 2026-11-15 | longshot v3.0.0-rc.4-11-ga00c485c-dirty (source 0b226774): main a00c485c with this change’s two fixes in the working tree (not committed when run) and a local counter of dark frames (not in the change); built with make_game.py |
Windows 11 Pro 26200, no window (headless, scripted input) | RTX 4070 SUPER 12 GB, NVIDIA 617.42 | Ryzen 9 9950X3D (16 cores, SMT off), 32 GB | 2560x1440, DLAA, no frame generation, uncapped (HOS_FPS_MAX=1000); RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32, RX_VK_LOAD=create, RX_VK_TEST_ALLOWED_MB=10200; the submitting thread with both changes off, main’s behaviour (RX_VK_ASYNC=0 RX_VK_MET_WAIT_CAPTURE=0 RX_VK_HOLD_GPU_WAIT=0) |
Castle Town’s south gate, save set “First Step into the game world Town”: out to the drawbridge and back through the gate several times on a wall-clock script (a few loading screens), night at 90 s, 140 s; a fourth round, after the change was moved onto main a00c485c, taken in turn with the two other sides | run sa4 (round 4, the one on main a00c485c): 37 dark-Link frames (frames drawn with an empty light grid, counted in the build) in 21137 frames; 2 hold timeouts (50 ms in the log’s order line); mean 141.3 fps over the run, loads included; 48 frames over 50 ms; 5760 captures at a CPU event’s wait, 0 of them at a wait met on arrival; stopped queues held for a GPU wait 0 times |
| 2026-11-15 | longshot v3.0.0-rc.4-1-ge090cc48-dirty (source 09e03930): main e090cc48 with this change’s two fixes in the working tree (not committed when run) and a local counter of dark frames (not in the change); built with make_game.py |
Windows 11 Pro 26200, no window (headless, scripted input) | RTX 4070 SUPER 12 GB, NVIDIA 617.42 | Ryzen 9 9950X3D (16 cores, SMT off), 32 GB | 2560x1440, DLAA, no frame generation, uncapped (HOS_FPS_MAX=1000); RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32, RX_VK_LOAD=create, RX_VK_TEST_ALLOWED_MB=10200; the submitting thread with both changes off, main’s behaviour (RX_VK_ASYNC=0 RX_VK_MET_WAIT_CAPTURE=0 RX_VK_HOLD_GPU_WAIT=0) |
Castle Town’s south gate, save set “First Step into the game world Town”: out to the drawbridge and back through the gate several times on a wall-clock script (a few loading screens), night at 90 s, 140 s; one of three rounds taken in turn with the two other sides | run sa3 (round 3 of 3): 21 dark-Link frames (frames drawn with an empty light grid, counted in the build) in 19416 frames; 1 hold timeouts (50 ms in the log’s order line); mean 123.2 fps over the run, loads included; 42 frames over 50 ms; 6060 captures at a CPU event’s wait, 0 of them at a wait met on arrival; stopped queues held for a GPU wait 0 times |
| 2026-11-15 | longshot v3.0.0-rc.4-1-ge090cc48-dirty (source 09e03930): main e090cc48 with this change’s two fixes in the working tree (not committed when run) and a local counter of dark frames (not in the change); built with make_game.py |
Windows 11 Pro 26200, no window (headless, scripted input) | RTX 4070 SUPER 12 GB, NVIDIA 617.42 | Ryzen 9 9950X3D (16 cores, SMT off), 32 GB | 2560x1440, DLAA, no frame generation, uncapped (HOS_FPS_MAX=1000); RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32, RX_VK_LOAD=create, RX_VK_TEST_ALLOWED_MB=10200; the submitting thread with both changes off, main’s behaviour (RX_VK_ASYNC=0 RX_VK_MET_WAIT_CAPTURE=0 RX_VK_HOLD_GPU_WAIT=0) |
Castle Town’s south gate, save set “First Step into the game world Town”: out to the drawbridge and back through the gate several times on a wall-clock script (a few loading screens), night at 90 s, 140 s; one of three rounds taken in turn with the two other sides | run sa2 (round 2 of 3): 28 dark-Link frames (frames drawn with an empty light grid, counted in the build) in 19660 frames; 2 hold timeouts (50 ms in the log’s order line); mean 125.1 fps over the run, loads included; 37 frames over 50 ms; 6171 captures at a CPU event’s wait, 0 of them at a wait met on arrival; stopped queues held for a GPU wait 0 times |
| 2026-11-15 | longshot v3.0.0-rc.4-1-ge090cc48-dirty (source 09e03930): main e090cc48 with this change’s two fixes in the working tree (not committed when run) and a local counter of dark frames (not in the change); built with make_game.py |
Windows 11 Pro 26200, no window (headless, scripted input) | RTX 4070 SUPER 12 GB, NVIDIA 617.42 | Ryzen 9 9950X3D (16 cores, SMT off), 32 GB | 2560x1440, DLAA, no frame generation, uncapped (HOS_FPS_MAX=1000); RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32, RX_VK_LOAD=create, RX_VK_TEST_ALLOWED_MB=10200; the submitting thread with both changes off, main’s behaviour (RX_VK_ASYNC=0 RX_VK_MET_WAIT_CAPTURE=0 RX_VK_HOLD_GPU_WAIT=0) |
Castle Town’s south gate, save set “First Step into the game world Town”: out to the drawbridge and back through the gate several times on a wall-clock script (a few loading screens), night at 90 s, 140 s; one of three rounds taken in turn with the two other sides | run sa1 (round 1 of 3): 40 dark-Link frames (frames drawn with an empty light grid, counted in the build) in 19149 frames; 3 hold timeouts (50 ms in the log’s order line); mean 121.3 fps over the run, loads included; 35 frames over 50 ms; 6080 captures at a CPU event’s wait, 0 of them at a wait met on arrival; stopped queues held for a GPU wait 0 times |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT=soa | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; order aos, indexed, indexed, aos, then soa, soa (soa not interleaved with aos; the interleaved pairs are runs fip-kH-*) | Run fi-soap-2: standing 11.53 fps, GPU time 53.1 ms, GPU busy 76%; walking 11.93 fps / GPU 43.5 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT=soa | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; order aos, indexed, indexed, aos, then soa, soa (soa not interleaved with aos; the interleaved pairs are runs fip-kH-*) | Run fi-soap-1: standing 13.81 fps, GPU time 43.5 ms, GPU busy 65%; walking 11.45 fps / GPU 44.3 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); RX_VK_PASS_TIMING=2 and RX_VK_FALLBACK_STATS=1; RX_VK_MESH_FB_LAYOUT=soa | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; order aos, indexed, soa, indexed, aos, soa | Run fi-soa-2: mesh 8.78 ms a frame, draw of the records 5.08 ms, before its draws 11.88 ms; hot mesh 1.93 / 1.76 / 0.73 / 0.75 ms; 1.084 M records, 54.9 MB a frame; overflow 0; profiled standing 10.41 fps / GPU 55.3 ms, walking 11.37 fps / 44.7 ms; pictures as in aos by eye. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); RX_VK_PASS_TIMING=2 and RX_VK_FALLBACK_STATS=1; RX_VK_MESH_FB_LAYOUT=soa | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; order aos, indexed, soa, indexed, aos, soa | Run fi-soa-1: mesh 7.91 ms a frame, draw of the records 4.63 ms, before its draws 10.57 ms; hot mesh 1.81 / 1.49 / 0.80 / 0.77 ms; 1.084 M records, 54.6 MB a frame; overflow 0; profiled standing 12.27 fps / GPU 48.3 ms, walking 11.41 fps / 45.3 ms; terrain, grass, foliage and shadows look as in aos (the scene animates, so not compared pixel by pixel). Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x800, render resolution 100%, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); RX_VK_PASS_TIMING=2 and RX_VK_FALLBACK_STATS=1; RX_VK_MESH_FB_LAYOUT=soa | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; order aos stats, soa stats, soa plain, aos plain | Run fi-kN-soa-s: mesh 7.17 ms a frame, draw of the records 6.41 ms, before its draws 9.96 ms; hot mesh 1.71 / 1.44 / 0.41 / 0.51 ms; 1.007 M records, 50.4 MB a frame; overflow 0; profiled standing 8.63 fps / GPU 70.9 ms, walking 9.06 fps / 70.3 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x800, render resolution 100%, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT=soa | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; order aos stats, soa stats, soa plain, aos plain | Run fi-kN-soa-p: standing 8.29 fps / GPU time 74.2 ms, walking 8.98 fps / 70.0 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x800, render resolution 100%, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); RX_VK_PASS_TIMING=2 and RX_VK_FALLBACK_STATS=1; RX_VK_MESH_FB_LAYOUT unset (aos) | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; order aos stats, soa stats, soa plain, aos plain | Run fi-kN-aos-s: mesh 14.00 ms a frame, draw of the records 6.53 ms, before its draws 16.50 ms; hot mesh 43ec 2.21 / d8e5 2.35 / 3b8c 0.89 / 89ca 0.89 ms; 1.003 M records, 49.8 MB a frame; overflow 0; profiled standing 8.41 fps / GPU 73.1 ms, walking 8.60 fps / 72.3 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x800, render resolution 100%, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT unset (aos) | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; order aos stats, soa stats, soa plain, aos plain | Run fi-kN-aos-p: standing 8.89 fps / GPU time 70.0 ms, walking 8.79 fps / 72.2 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x800, render resolution 100%, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); RX_VK_PASS_TIMING=2 and RX_VK_FALLBACK_STATS=1; RX_VK_MESH_FB_LAYOUT=soa | the main outdoor area at the Lost Woods bridge, child, reached by teleport from the starting-area save, 60 s warm-up then 120 s standing; order aos stats, soa stats, soa plain, aos plain | Run fi-fN-soa-s: the fallback’s mesh dispatches 4.37 ms a frame (-47% against fi-fN-aos-s); overflow 0. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x800, render resolution 100%, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT=soa | the main outdoor area at the Lost Woods bridge, child, reached by teleport from the starting-area save, 60 s warm-up then 120 s standing; order aos stats, soa stats, soa plain, aos plain | Run fi-fN-soa-p: 15.87 fps, p99 72.0 ms, GPU time 41.2 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x800, render resolution 100%, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); RX_VK_PASS_TIMING=2 and RX_VK_FALLBACK_STATS=1; RX_VK_MESH_FB_LAYOUT unset (aos) | the main outdoor area at the Lost Woods bridge, child, reached by teleport from the starting-area save, 60 s warm-up then 120 s standing; order aos stats, soa stats, soa plain, aos plain | Run fi-fN-aos-s: the fallback’s mesh dispatches 8.24 ms a frame; overflow 0. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x800, render resolution 100%, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT unset (aos) | the main outdoor area at the Lost Woods bridge, child, reached by teleport from the starting-area save, 60 s warm-up then 120 s standing; order aos stats, soa stats, soa plain, aos plain | Run fi-fN-aos-p: 15.01 fps, p99 73.4 ms, GPU time 44.9 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); RX_VK_PASS_TIMING=2 and RX_VK_FALLBACK_STATS=1; RX_VK_MESH_FB_LAYOUT=soa | the main outdoor area at the Lost Woods bridge, child, reached by teleport from the starting-area save, 60 s warm-up then 120 s standing; order aos stats, soa stats, soa plain, aos plain | Run fi-fH-soa-s: mesh 4.38 ms a frame, draw of the records 2.11 ms, before its draws 5.61 ms; hot mesh 1.52 / 0.97 / 0.08 / 0.31 ms; 0.917 M records, 37.5 MB a frame; overflow 0; profiled 21.63 fps / GPU 23.5 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT=soa | the main outdoor area at the Lost Woods bridge, child, reached by teleport from the starting-area save, 60 s warm-up then 120 s standing; order aos stats, soa stats, soa plain, aos plain | Run fi-fH-soa-p: 22.39 fps, GPU time 23.0 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); RX_VK_PASS_TIMING=2 and RX_VK_FALLBACK_STATS=1; RX_VK_MESH_FB_LAYOUT unset (aos) | the main outdoor area at the Lost Woods bridge, child, reached by teleport from the starting-area save, 60 s warm-up then 120 s standing; order aos stats, soa stats, soa plain, aos plain | Run fi-fH-aos-s: mesh 8.42 ms a frame, draw of the records 2.15 ms, before its draws 9.63 ms; hot mesh 43ec 2.05 / d8e5 1.44 / 3b8c 0.10 / 89ca 0.42 ms; 0.917 M records, 37.5 MB a frame; overflow 0; profiled 19.86 fps / GPU 27.5 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT unset (aos) | the main outdoor area at the Lost Woods bridge, child, reached by teleport from the starting-area save, 60 s warm-up then 120 s standing; order aos stats, soa stats, soa plain, aos plain | Run fi-fH-aos-p: 20.54 fps, GPU time 26.9 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT unset (aos) | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; order aos, indexed, indexed, aos, then soa, soa (soa not interleaved with aos; the interleaved pairs are runs fip-kH-*) | Run fi-aosp-2: standing 9.96 fps, frame p50 92.5 / p99 198.1 ms, GPU time 58.9 ms; walking 9.91 fps / GPU 57.5 ms; GPU busy 80%. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); no timing or stats switches (fps arm); RX_VK_MESH_FB_LAYOUT unset (aos) | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; order aos, indexed, indexed, aos, then soa, soa (soa not interleaved with aos; the interleaved pairs are runs fip-kH-*) | Run fi-aosp-1: standing 10.23 fps, frame p50 84.6 / p99 201.3 ms, GPU time 55.9 ms; walking 10.86 fps / GPU 52.1 ms; GPU busy 80%. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); RX_VK_PASS_TIMING=2 and RX_VK_FALLBACK_STATS=1; RX_VK_MESH_FB_LAYOUT unset (aos) | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; order aos, indexed, soa, indexed, aos, soa | Run fi-aos-2: mesh 16.25 ms a frame, draw of the records 4.50 ms, before its draws 18.86 ms; hot mesh 2.35 / 2.06 / 1.48 / 1.26 ms; 1.083 M records, 54.8 MB a frame; overflow 0; profiled standing 11.18 fps / GPU 55.8 ms, walking 10.70 fps / 50.0 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 UTC | fallback-records-indexed 94a0586d (the four-commit form of this PR; its soa translation is byte for byte this PR’s), renderer-only dev_packs pack on main 9c9d69ed with the fallback probe’s timing and stats counters (RX_VK_PASS_TIMING=2 parts, RX_VK_FALLBACK_STATS=1; pack tree 4d5ef359); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1) | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock auto clocks | Zen 2 4C/8T, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, Quality preset Custom, frame generation off; empty program and pipeline caches with RX_VK_PRECOMPILE=0 (each layout translated on demand, the first report ignored); RX_VK_PASS_TIMING=2 and RX_VK_FALLBACK_STATS=1; RX_VK_MESH_FB_LAYOUT unset (aos) | the starting forest, a child save standing beside Saria (about 2,200 draws a frame) for 120 s, then 3 scripted walk cycles; order aos, indexed, soa, indexed, aos, soa | Run fi-aos-1: the fallback’s mesh dispatches 17.81 ms a frame, draw of the records 5.02 ms, fallback before its draws 20.76 ms; hot programs’ mesh 43ec 2.38 / d8e5 2.27 / 3b8c 1.50 / 89ca 1.31 ms; 1.085 M records, 54.9 MB written a frame; overflow 0; profiled standing 10.34 fps / GPU 59.9 ms, walking 11.14 fps / 49.1 ms. Exit 0, no device loss, no amdgpu kernel lines. |
| 2026-11-15 | main 7ff5256e against the combined build of main 7ff5256e with https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/550 to https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/558 merged (https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/558 @ 6fb80632), x86-64-v3, release play profile, LLVM 23.1.5 | SteamOS 3.9.5 (build 20261023.100), kernel 7.2.7-valve2, headless (HOS_HEADLESS=1) over SSH |
AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel git-5c65d4fc31 | Steam Deck OLED, AMD Custom APU 0932, 16 GB | logs/ not writable, XDG_DATA_HOME an empty test folder |
the start into the intro, once per build in a pseudo-terminal and once through systemd-cat (stderr the journal’s socket), closed through the control bridge |
Test request T-ZH5M: through systemd-cat main’s start-up messages went nowhere (0 lines) and https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/558’s all reached the journal (630 lines); from a terminal both showed them there (603 / 595 lines); the game ran in all four starts |
| 2026-11-15 | brett-d/legendary-broccoli 52a61f40 merged into Longshot main 9c9d69ed (local merge 0e4446d7), renderer built via scripts/dev_packs.py on a base of main 9c9d69ed, interleaved with main 9c9d69ed |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2), headless (HOS_HEADLESS=1) |
Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), stock GPU clock | Steam Deck OLED APU, 16 GB RAM (15.1 GB usable), 8.2 GB swap | Output 1280x800, Render resolution 100% (1280x800), Quality preset Custom, frame generation off, Texture budget Auto (2560 MB), default precompile on (RX_VK_LOAD=start) |
Child starting-area save, the main outdoor area teleport, turning camera (3 cycles of 420 game frames, 2 interleaved runs per arm) and 45 s standing in the main outdoor area (2 interleaved runs per arm; test request T-5P52) | Turning-camera route (2 warm runs each): main ran out of swap in both runs (SwapFree minimum 0 and 0 MB, VmSwap peak 4,044 and 4,029 MB, MemAvailable minimum 237 and 319 MiB, PSI full avg10 35% and 38%, fps 13.94 and 13.70); this PR let go of 38,078 shader modules in each run and kept SwapFree >= 2,267 and 2,176 MB (VmSwap peak 1,315 and 1,181 MB, MemAvailable minimum 608 and 936 MiB, PSI full avg10 17% and 17%, fps 14.40 and 14.02). Standing in the main outdoor area (2 warm runs each): VmSwap peak 3,966 and 3,939 MB (main) vs 133 and 633 MB (PR), SwapFree minimum 785 and 1,455 MB (main) vs 5,960 and 4,729 MB (PR), MemAvailable minimum 359 and 359 MiB (main) vs 538 and 1,037 MiB (PR), fps 14.69 and 14.93 (main) vs 14.96 and 14.95 (PR). Caches kept across update, 0 guard trips, 0 device losses |
| 2026-11-15 | local integration vrs-verify bfc69759: main aa95d35f with this PR (radv-vrs-key-switch), driver-cache-diagnostics and precompile-save-before-title, a game pack built via scripts/dev_packs.py on a base of main 9c9d69ed; the same pack for every start |
SteamOS 3.9.5 (20261023.100, kernel 7.2.7-valve2), headless | Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel (git-5c65d4fc31), stock GPU clock | Steam Deck OLED APU, 16 GB RAM | Defaults, the precompile at its default; scratch game folders, each start with a fresh copy of the same warm caches (32,438 noted graphics pipelines); arm b with RADV_FORCE_VRS_CONFIG_FILE naming a file holding 1x1, as Game Mode’s /tmp/radv_vrs.<random> does |
The start-up to the title, 4 starts interleaved a, b, b, a (run deck01): a without the variable, b with it | b: the variable dropped (vk: RADV_FORCE_VRS_CONFIG_FILE=<scratch>/vrs/vrs-1x1.txt dropped: ...), nothing compiled: precompile 18,923 and 12,481 ms, title in 51 and 45 s. a: precompile 17,379 and 20,670 ms, title in 52 and 54 s. Without this PR (an earlier build of main aa95d35f with other open PRs, the same caches) the b start compiled at about 20 pipelines a second (noted pipelines: 8932 of 38345 pipelines in 130 s, stopped at 150 s; 25-30 minutes for all). No warning about the pipelines’ keys in a or b |
| 2026-11-15 | main 7ff5256e against main with https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/556’s renderer from acd34022, x86-64-v3, release play profile, LLVM 23.1.5 | SteamOS 3.9.5 (build 20261023.100), kernel 7.2.7-valve2, Game Mode, docked | AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel git-5c65d4fc31 | Steam Deck OLED, AMD Custom APU 0932, 16 GB | main’s defaults on that Deck: Auto preset, output 1280x720, render 50%, 2.2 GB VRAM budget, frame generation off; Upscaler switched to AMD FSR 3.1 and back through the menu’s setter | the starting forest standing: FSR 3.1 for 10 s, back to the game’s own, the vk: VRAM: line after the release line; four runs interleaved https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/556, main, main, https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/556 |
Test request T-UH7H: after leaving FSR 3.1 the renderer held 1,107 / 1,076 MB besides textures with https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/556 against 1,154 / 1,177 MB on main (falling 115 / 102 MB within the run against -29 / +29); used minus free the same in both (3,585 to 3,607 MB, the textures took the space); FSR 3.1 again made its images and drew right; no fault |
| 2026-11-15 | longshot v3.0.0-rc.4-145-g6863feb4 (fix/fg-provider-row 6863feb4 on main 1764fed2), a dev runtime on the same base as main’s | Windows 11 Pro, fullscreen, 5120x2160 at 165 Hz | RTX 5090 32 GB, NVIDIA 621.24 | Ryzen 7 9800X3D, 64 GB | DLAA (preset M), Frame generation 3x, Frame generation by DLSS FG (the default), V-Sync On (FIFO), Reflex On + Boost; one options file and save for all four runs | Driven by the control bridge (scripts/longshot_ctl.py): the title screen for 10 s, A until play, 60 s standing still in Market_Child, PLUS for 5 s, quit; the run’s log presents, fps and DLSS-G 600-present lines |
NGX answered at the start-up screen; title screen: the game’s own 164.8 fps, 0 of 600 by DLSS-G, no made frames; the row “DLSS FG” at the first screen, the title screen and in play (bridge); the game went into play by itself with no bridge input, at another spot than main2’s: 600 of 600 by DLSS-G, the game’s own 42.1 fps median over 60 s (3167 draws a frame, not comparable with main2); the pause screen 165 fps, DLSS FG paused (run pr2) |
| 2026-11-15 | longshot v3.0.0-rc.4-145-g6863feb4 (fix/fg-provider-row 6863feb4 on main 1764fed2), a dev runtime on the same base as main’s | Windows 11 Pro, fullscreen, 5120x2160 at 165 Hz | RTX 5090 32 GB, NVIDIA 621.24 | Ryzen 7 9800X3D, 64 GB | DLAA (preset M), Frame generation 3x, Frame generation by DLSS FG (the default), V-Sync On (FIFO), Reflex On + Boost; one options file and save for all four runs | The title screen, then the save in play in Market_Child and the game’s pause screen, by hand; longshot_log.txt’s presents, fps and DLSS-G 600-present lines |
NGX answered at the start-up screen; title screen: the game’s own 164.8 fps, 0 of 600 by DLSS-G, no made frames, the row and counter as expected (“DLSS FG”, “DLSS FG paused”); in play 600 of 600 by DLSS-G, the game’s own 47.2 fps median over about 60 s; the pause screen about 160 fps, DLSS FG paused (run pr1) |
| 2026-11-15 | longshot v3.0.0-rc.4-143-g1764fed2 (main), a dev runtime on one base with the PR’s | Windows 11 Pro, fullscreen, 5120x2160 at 165 Hz | RTX 5090 32 GB, NVIDIA 621.24 | Ryzen 7 9800X3D, 64 GB | DLAA (preset M), Frame generation 3x, Frame generation by Auto (the default), V-Sync On (FIFO), Reflex On + Boost; one options file and save for all four runs | Driven by the control bridge (scripts/longshot_ctl.py): the title screen for 10 s, A until play, 60 s standing still in Market_Child, PLUS for 5 s, quit; the run’s log presents, fps and DLSS-G 600-present lines |
Title screen: the game’s own 54.6 fps, 0 of 600 by DLSS-G, the built-in one made the rest; the row “Auto: Built-in FG” (bridge); in play the row “Auto: DLSS FG”, 600 of 600 by DLSS-G, the game’s own 45.3 fps median over 60 s standing (2860 draws a frame); PLUS came during a game event: no pause screen (run main2) |
| 2026-11-15 | longshot v3.0.0-rc.4-143-g1764fed2 (main), a dev runtime on one base with the PR’s | Windows 11 Pro, fullscreen, 5120x2160 at 165 Hz | RTX 5090 32 GB, NVIDIA 621.24 | Ryzen 7 9800X3D, 64 GB | DLAA (preset M), Frame generation 3x, Frame generation by Auto (the default), V-Sync On (FIFO), Reflex On + Boost; one options file and save for all four runs | The title screen, then the save in play in Market_Child, by hand; longshot_log.txt’s presents, fps and DLSS-G 600-present lines |
Title screen: the game’s own 54.5 fps, 0 of 600 by DLSS-G, the built-in one made the rest; the row and counter as expected for the bug (“Auto: Built-in FG”); in play 600 of 600 by DLSS-G, the game’s own 47.0 fps median over about 17 s (run main1) |
| 2026-11-15 UTC | early-reads-port’s gpu/ diff (the early reads from sheep/BunnyHood#2 and #303, off by default) on Longshot main 9c9d69ed with #443’s vk_replay.cpp change, as a renderer-only scripts/dev_packs.py pack (longshot_vk.so sha256 f34da1d5…51928; the game binary is main 9c9d69ed’s); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1), scripted input | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock automatic GPU clock | AMD Zen 2 4C/8T APU, 16 GB shared | output 1280x800, render resolution 100%, the game’s own upscaler, quality preset Custom, frame generation off, render thread off (Auto); warm shader caches copied fresh, RX_VK_PRECOMPILE=0, no longshot_env.txt; RX_VK_READBACK_EARLY unset (off, main’s path) against =1 (on), interleaved off, on, on, off |
the main outdoor area at the Lost Woods bridge, teleported from a child starting-area save, 20 s settle then 45 s standing (runs e2-off-fN-1, e2-on-fN-1, e2-on-fN-2, e2-off-fN-2) | 14.97 and 14.99 fps off against 18.45 and 18.48 on: median 14.98 to 18.47 fps (+23.3%); p50 66.6 and 66.5 ms against 51.7 and 52.0; GPU time 44.1-44.9 ms a frame in both; the host’s GPU waits 39.0 against 22.2-22.6 ms a frame; GPU busy 74-76% against 88-89%; memory sync 0.8 ms a frame on; exit 0, no device loss. |
| 2026-11-15 UTC | early-reads-port’s gpu/ diff (the early reads from sheep/BunnyHood#2 and #303, off by default) on Longshot main 9c9d69ed with #443’s vk_replay.cpp change, as a renderer-only scripts/dev_packs.py pack (longshot_vk.so sha256 f34da1d5…51928; the game binary is main 9c9d69ed’s); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1), scripted input | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock automatic GPU clock | AMD Zen 2 4C/8T APU, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, quality preset Custom, frame generation off, render thread off (Auto); warm shader caches copied fresh, RX_VK_PRECOMPILE=0, no longshot_env.txt; RX_VK_READBACK_EARLY unset (off, main’s path) against =1 (on), interleaved off, on, on, off after a warm-up |
the starting forest, a child save outside the player character’s house beside Saria (about 2,200 draws a frame), 45 s standing (runs e2-off-kH-1, e2-on-kH-1, e2-on-kH-2, e2-off-kH-2) | 11.82 and 10.98 fps off against 14.11 and 13.67 on: median 11.40 to 13.89 fps (+21.8%); p99 127.3 and 157.9 ms off against 119.0 and 126.5 on; frames over 100 ms 99 and 146 against 36 and 44; GPU time 52-55 ms a frame in both; the host’s GPU waits 48.5-51.0 against 22.6-24.1 ms a frame; GPU busy 75-77% against 86-87%; memory sync 1.8-1.9 ms a frame on; exit 0, no device loss. |
| 2026-11-15 UTC | early-reads-port’s gpu/ diff (the early reads from sheep/BunnyHood#2 and #303, off by default) on Longshot main 9c9d69ed with #443’s vk_replay.cpp change, as a renderer-only scripts/dev_packs.py pack (longshot_vk.so sha256 f34da1d5…51928; the game binary is main 9c9d69ed’s); tester zeldaootr |
SteamOS 3.9.5 (20261023.100), kernel 7.2.7-valve2, headless over ssh (HOS_HEADLESS=1), scripted input | Steam Deck OLED, AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel, stock automatic GPU clock | AMD Zen 2 4C/8T APU, 16 GB shared | output 1280x720, render resolution 50%, the game’s own upscaler, quality preset Custom, frame generation off, render thread off (Auto); warm shader caches copied fresh, RX_VK_PRECOMPILE=0, no longshot_env.txt; RX_VK_READBACK_EARLY=1 RX_VK_LOG_EARLY_CHECK=4 RX_VK_LOG_READBACK_WAITS=1 (the byte probe: one early read in four checked against main’s wait) |
the starting forest, a child save outside the player character’s house beside Saria: 30 s standing then 12 scripted walk-and-pan cycles (run e2-probe-starting-area-1); then the starting forest into the first dungeon and back twice, camera pans at each arrival (run e2-probe-forest-2) | Byte probe: 2,300 (starting-area walking) and 1,900 (starting-area and the first dungeon) reads checked, 4,200 in all; 0 changed by main’s release on pages the CPU did not write (0 missed), 0 lost, 0 ambiguous before main’s release; the early-reads log lines confirmed engaged; exit 0, no device loss. |
| 2026-11-15 | audio runtime 05bc368136f26c9a8ad0f34ccc7e5d7e11b979ff; main a0be15b318936245b80bbf57286826af9c549152 | Windows 10 Pro 19045; local interactive desktop | RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT; 63.9 GiB usable RAM | Dedicated audio test install and independent user data; named virtual output; Mono and microphone controls | windows-play05; owner-operated playback/menu check, disable and re-enable selected endpoint in Windows Sound settings | Owner reports all checks pass: fallback, retained selection, restored output, playback, Mono and microphone selection; virtual-endpoint disappearance only, not physical USB hotplug; old translator cache rejected and startup shader preparation completed in 166.6 s |
| 2026-11-15 | clean audio topic ca70a6b82831f78c8147e558b1d59fd55e94b827 | Windows 10 Pro build 19045; fullscreen desktop session | NVIDIA RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT, 16 cores/32 threads; 63.9 GiB usable RAM | Custom preset; game output 2048x1152, graphics scale 1; 2560x1440 swapchain; Catmull-Rom, render resolution 100%; frame generation Off; V-sync Auto/mailbox; HOS_FPS_MAX=1000; FPS counter Breakdown; selector Default, stereo output, audio enabled | windows-ab-topic02; stationary the village child/night from an independent copy of the same save; matching position and camera; separate 45 s warm-ups of both builds, then 30 s settling and 60 s recording per run; second measured topic in main/topic/main/topic | Median native 120-present FPS samples: 51.2 FPS; no new pipelines or shader translations during the timed window; actor snapshots remained fresh and stationary; two-run topic median 51.05 FPS versus main 51.45 FPS, about 0.8 percent lower on topic and within observed run-to-run spread; two pairs in one scene do not rule out smaller costs |
| 2026-11-15 | clean audio topic ca70a6b82831f78c8147e558b1d59fd55e94b827 | Windows 10 Pro build 19045; fullscreen desktop session | NVIDIA RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT, 16 cores/32 threads; 63.9 GiB usable RAM | Custom preset; game output 2048x1152, graphics scale 1; 2560x1440 swapchain; Catmull-Rom, render resolution 100%; frame generation Off; V-sync Auto/mailbox; HOS_FPS_MAX=1000; FPS counter Breakdown; selector Default, stereo output, audio enabled | windows-ab-topic01; stationary the village child/night from an independent copy of the same save; matching position and camera; separate 45 s warm-ups of both builds, then 30 s settling and 60 s recording per run; first measured topic in main/topic/main/topic | Median native 120-present FPS samples: 50.9 FPS; no new pipelines or shader translations during the timed window; actor snapshots remained fresh and stationary; combined two-pair verdict recorded separately; not a frame-generation rate or proof of zero overhead |
| 2026-11-15 | clean main 3dec76818058c2c3eb66b4aca302962b79d96305 | Windows 10 Pro build 19045; fullscreen desktop session | NVIDIA RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT, 16 cores/32 threads; 63.9 GiB usable RAM | Custom preset; game output 2048x1152, graphics scale 1; 2560x1440 swapchain; Catmull-Rom, render resolution 100%; frame generation Off; V-sync Auto/mailbox; HOS_FPS_MAX=1000; FPS counter Breakdown; default stereo output, audio enabled | windows-ab-main02; stationary the village child/night from an independent copy of the same save; matching position and camera; separate 45 s warm-ups of both builds, then 30 s settling and 60 s recording per run; second measured main in main/topic/main/topic | Median native 120-present FPS samples: 51.2 FPS; no new pipelines or shader translations during the timed window; actor snapshots remained fresh and stationary; two-run main median 51.45 FPS versus topic 51.05 FPS, about 0.8 percent lower on topic and within observed run-to-run spread; two pairs in one scene do not rule out smaller costs |
| 2026-11-15 | clean main 3dec76818058c2c3eb66b4aca302962b79d96305 | Windows 10 Pro build 19045; fullscreen desktop session | NVIDIA RTX 3080; driver 32.0.16.2124 | Ryzen 9 5900XT, 16 cores/32 threads; 63.9 GiB usable RAM | Custom preset; game output 2048x1152, graphics scale 1; 2560x1440 swapchain; Catmull-Rom, render resolution 100%; frame generation Off; V-sync Auto/mailbox; HOS_FPS_MAX=1000; FPS counter Breakdown; default stereo output, audio enabled | windows-ab-main01; stationary the village child/night from an independent copy of the same save; matching position and camera; separate 45 s warm-ups of both builds, then 30 s settling and 60 s recording per run; first measured main in main/topic/main/topic | Median native 120-present FPS samples: 51.7 FPS; no new pipelines or shader translations during the timed window; actor snapshots remained fresh and stationary; combined two-pair verdict recorded separately; not a frame-generation rate or proof of zero overhead |
| 2026-11-15 | feat/audio-output-selector 5e0891c1e132855dc0898b199109629d1aecef73 on main aa95d35f; source f7b73e9a | Ubuntu 26.04.1, kernel 7.0.0-43-generic; local Wayland desktop, SDL2 | RX Vega M GH 4 GB, RADV Mesa 26.0.8, Vulkan 1.4 | Intel i7-8809G, 4 cores/8 threads, 32 GB RAM | System-default stereo playback; prepared caches; final logged swapchain 1853x1131, mailbox; other scene settings not fully recorded | linux-play01; startup and Audio-menu functional smoke, precise gameplay area not recorded; owner operated controls; 2m52.969s captured desktop session, then window closed | Owner reported success for requested playback, moved audio controls and microphone-selector smoke checks, then explicitly clarified Mono was not tested in this run; logs show cache preload 4.890s, default stereo output, input-stream initialization and 999 played blocks with nonzero samples in first 10 s; no fatal/audio-device error identified; deliberate GPU network omissions logged; no named Linux output support, independent end-to-end mic capture proof, text-input test or paired FPS measurement |
| 2026-11-15 | feat/audio-output-selector 5e0891c1e132855dc0898b199109629d1aecef73 on main aa95d35f; source f7b73e9a | Ubuntu 26.04.1, kernel 7.0.0-43-generic; local Wayland desktop | RX Vega M GH 4 GB, RADV Mesa 26.0.8 | Intel i7-8809G, 4 cores/8 threads, 32 GB RAM | Owner manually relaunched the prepared native Linux build and tested Mono in Audio; remaining settings not recorded | linux-mono01; separate owner-operated Mono follow-up after linux-play01; exact scene and duration unknown | Owner confirmed Mono was working; this is human-reported functional evidence, with no captured follow-up log, PCM-channel comparison, screenshot, restart-persistence proof or performance measurement; does not validate latest main or Windows named-device routing |
| 2026-11-15 | API22 base82886e89 plus scheduling, cache-input, progress/readiness fixes, d41dfcc5 | Windows 11 Home N26300; native preparation | RX 9070 XT16GB; AMD26.11.1 | Ryzen 7 5700X3D;64GiB | Default longshot.exe —prepare-shaders —compile-shaders all; normal watchdog and memory guards; matching renderer PGO | Preparation only after driver change; no scene drawn | Exit0; shader stage complete;24626 noted graphics and9687 compute pipelines,0/0 not made and0 missing program variants. Actual GPU cache saved. Combined integration sanity, not an isolated Most-to-All timing or gameplay result |
| 2026-11-15 | API22 base82886e89 plus progress/readiness fixes, b7b2fbaa | Windows 11 Home N26300; native preparation | RX 9070 XT16GB; AMD26.11.1 | Ryzen 7 5700X3D;64GiB | Default longshot.exe —prepare-shaders —compile-shaders all; normal watchdog and memory guards; matching renderer PGO | Preparation only after driver change; no scene drawn | Exit0; shader stage complete;24626 noted graphics and9687 compute pipelines,0/0 not made and0 missing program variants. Actual GPU cache saved. Combined integration sanity, not an isolated Most-to-All timing or gameplay result |
| 2026-11-14 | b46daa73 (Intel XeSS Super Resolution, on main cf605c85) with three changes not merged yet also in: the Battlemage float-add fix, the uniform snapshot on Arc discrete cards and the split-before-draw rule; the renderer built from source into a nightly-80-7315957c game folder; libxess.dll from Intel’s XeSS SDK 3.0.2 (reports XeSS 2.0.2) |
Windows 11 Pro 10.0.29704, 2048x1152 output | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | Intel XeSS Native AA (RX_VK_XESS=2, 2048x1152); render size 1366x768 (67%) but for Native AA; frame generation off, uncapped (HOS_FPS_MAX=1000) |
the starting forest (child), the spawn spot on the player character’s balcony after the opening, about 50 s standing still | fps over the last eight 5 s windows 33.1, 34.4, 34.2, 34.5, 34.2, 33.3, 33.7, 34.3, median 34.2; GPU time 15.4 ms a frame; the picture right (run X6) |
| 2026-11-14 | b46daa73 (Intel XeSS Super Resolution, on main cf605c85) with three changes not merged yet also in: the Battlemage float-add fix, the uniform snapshot on Arc discrete cards and the split-before-draw rule; the renderer built from source into a nightly-80-7315957c game folder; libxess.dll from Intel’s XeSS SDK 3.0.2 (reports XeSS 2.0.2) |
Windows 11 Pro 10.0.29704, 2048x1152 output | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | Intel XeSS (RX_VK_XESS=1, 1366x768 -> 2048x1152, Ultra Quality); render size 1366x768 (67%) but for Native AA; frame generation off, uncapped (HOS_FPS_MAX=1000) |
the starting forest (child), the spawn spot on the player character’s balcony after the opening, about 50 s standing still | fps over the last eight 5 s windows 35.3, 34.9, 35.5, 33.8, 31.0, 32.2, 33.6, 34.6, median 34.2; GPU time 13.4 ms a frame; the picture right (run X5) |
| 2026-11-14 | b46daa73 (Intel XeSS Super Resolution, on main cf605c85) with three changes not merged yet also in: the Battlemage float-add fix, the uniform snapshot on Arc discrete cards and the split-before-draw rule; the renderer built from source into a nightly-80-7315957c game folder; libxess.dll from Intel’s XeSS SDK 3.0.2 (reports XeSS 2.0.2) |
Windows 11 Pro 10.0.29704, 2048x1152 output | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | the game’s own upscaler (RX_VK_XESS=0, Intel’s library not loaded); render size 1366x768 (67%) but for Native AA; frame generation off, uncapped (HOS_FPS_MAX=1000) |
the starting forest (child), the spawn spot on the player character’s balcony after the opening, about 50 s standing still | fps over the last eight 5 s windows 37.7, 37.7, 38.3, 38.4, 37.8, 38.4, 38.2, 37.6, median 38.0; GPU time 11.6 ms a frame; the picture right (run X4) |
| 2026-11-14 | b46daa73 (Intel XeSS Super Resolution, on main cf605c85) with three changes not merged yet also in: the Battlemage float-add fix, the uniform snapshot on Arc discrete cards and the split-before-draw rule; the renderer built from source into a nightly-80-7315957c game folder; libxess.dll from Intel’s XeSS SDK 3.0.2 (reports XeSS 2.0.2) |
Windows 11 Pro 10.0.29704, 2048x1152 output | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | Intel XeSS Native AA (RX_VK_XESS=2, 2048x1152); render size 1366x768 (67%) but for Native AA; frame generation off, uncapped (HOS_FPS_MAX=1000) |
the starting forest (child), the spawn spot on the player character’s balcony after the opening, about 50 s standing still | fps over the last eight 5 s windows 34.0, 34.6, 34.3, 34.7, 33.9, 35.0, 34.1, 34.4, median 34.4; GPU time 15.1 ms a frame; the picture right (run X3) |
| 2026-11-14 | b46daa73 (Intel XeSS Super Resolution, on main cf605c85) with three changes not merged yet also in: the Battlemage float-add fix, the uniform snapshot on Arc discrete cards and the split-before-draw rule; the renderer built from source into a nightly-80-7315957c game folder; libxess.dll from Intel’s XeSS SDK 3.0.2 (reports XeSS 2.0.2) |
Windows 11 Pro 10.0.29704, 2048x1152 output | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | Intel XeSS (RX_VK_XESS=1, 1366x768 -> 2048x1152, Ultra Quality); render size 1366x768 (67%) but for Native AA; frame generation off, uncapped (HOS_FPS_MAX=1000) |
the starting forest (child), the spawn spot on the player character’s balcony after the opening, about 50 s standing still | fps over the last eight 5 s windows 35.1, 35.5, 35.3, 35.6, 35.3, 35.0, 35.7, 34.7, median 35.3; GPU time 13.3 ms a frame; the picture right (run X2) |
| 2026-11-14 | b46daa73 (Intel XeSS Super Resolution, on main cf605c85) with three changes not merged yet also in: the Battlemage float-add fix, the uniform snapshot on Arc discrete cards and the split-before-draw rule; the renderer built from source into a nightly-80-7315957c game folder; libxess.dll from Intel’s XeSS SDK 3.0.2 (reports XeSS 2.0.2) |
Windows 11 Pro 10.0.29704, 2048x1152 output | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | the game’s own upscaler (RX_VK_XESS=0, Intel’s library not loaded); render size 1366x768 (67%) but for Native AA; frame generation off, uncapped (HOS_FPS_MAX=1000) |
the starting forest (child), the spawn spot on the player character’s balcony after the opening, about 50 s standing still | fps over the last eight 5 s windows 35.7, 33.7, 32.4, 35.4, 35.3, 33.8, 35.8, 32.6, median 34.6; GPU time 11.7 ms a frame; the picture right (run X1) |
| 2026-11-14 | this PR d0340ee9; game-free Python check | Windows 11 build26300, native Python3.14.7, no desktop interaction | Not used; no game or GPU work | Ryzen7 9800X3D, 96GiB | One held-pipe timeout/recovery method; actual Windows named pipes | Authored tool fixtures; run02-pr, interleaved main/PR/main/PR; no game scene | 1 methods, 0 assertion/subtest failures, 0 errors, 0 skips; no performance claim; run02-pr |
| 2026-11-14 | main 9c9d69ed; game-free Python check | Windows 11 build26300, native Python3.14.7, no desktop interaction | Not used; no game or GPU work | Ryzen7 9800X3D, 96GiB | One held-pipe timeout/recovery method; actual Windows named pipes | Authored tool fixtures; run02-main, interleaved main/PR/main/PR; no game scene | 1 methods, 1 assertion/subtest failures, 0 errors, 0 skips; no performance claim; run02-main |
| 2026-11-14 | this PR d0340ee9; game-free Python check | Windows 11 build26300, native Python3.14.7, no desktop interaction | Not used; no game or GPU work | Ryzen7 9800X3D, 96GiB | One held-pipe timeout/recovery method; actual Windows named pipes | Authored tool fixtures; run01-pr, interleaved main/PR/main/PR; no game scene | 1 methods, 0 assertion/subtest failures, 0 errors, 0 skips; no performance claim; run01-pr |
| 2026-11-14 | main 9c9d69ed; game-free Python check | Windows 11 build26300, native Python3.14.7, no desktop interaction | Not used; no game or GPU work | Ryzen7 9800X3D, 96GiB | One held-pipe timeout/recovery method; actual Windows named pipes | Authored tool fixtures; run01-main, interleaved main/PR/main/PR; no game scene | 1 methods, 1 assertion/subtest failures, 0 errors, 0 skips; no performance claim; run01-main |
| 2026-11-14 | Game v3.0.0-rc.3-75-g9c9d69ed (source6b76ded3), matched API22 modules; control client d0340ee9 | Windows11 build26300, headless | RTX5090, NVIDIA617.42 | Ryzen7 9800X3D, 96GiB | Independent copied profile; audio0, DLSS0, framegen0, cap60; requested1280x720, bridge window1920x1080 | Title-menu status, frames1339/1354/1367/1382; four consecutive metadata-only connections; live01 | Four successful protocol1 replies; normal quit, native0/guard0; all23 protected hashes unchanged. Busy occurrence not established; no performance or gameplay-input claim; live01 |
| 2026-11-14 | main 0caf0618 against its renderer built with the play profile (RX_VK_PGO, build/vk-pgo a601a547), the same base otherwise |
CachyOS (Linux) | NVIDIA GeForce RTX 3090, 620.70 | Ryzen 9 7900X, 32 GB | 2560x1440, default options; low priority on a busy PC, default reserved core, no ReplayPin in the base | The Market from a save, standing then running in the square; full-length sessions interleaved control, profile, profile, control, control, profile (runs vkc-1 to vkc-6) | Median frame standing 13.53 / 13.49 / 13.70 ms without the profile, 12.96 / 13.37 / 12.94 ms with; running 15.89 / 16.13 / 15.73 ms without, 15.13 / 15.67 / 15.52 ms with (about -4 to -5 % and -3 %); a fourth control in a noisy stretch (vkc-8) 14.63 / 17.95 ms left out |
| 2026-11-14 | main 26a2eab003eeb18db69bcc41e1346dc4aec64b35 plus this change (the host/hos_thread.c change of the vcore starvation pull request, built alone on that commit: source stamp 6c3fbf7a, longshot sha256 2230b13d…), HOS_VCORE_STARVE_US=0 (the old rule); tester mooki; run identifier ab1-old-rule, starts 01-10 | Arch Linux 7.2.15-1-omarchy, windowless/offscreen | RTX 4090 24 GB, NVIDIA 615.78.08 | Core i9-13900K, 96 GB RAM | 1920x1080, Custom preset, render 100%, game upscaler and frame generation Off, uncapped, real-time speed, HOS_FPS_MAX=1000; shaders prepared | the starting forest (child), a copied test save loaded from the title to play, 10 starts interleaved with run01 (the same build, default), three starts to one hold, 5 s standing after play began; counts and log lines, not frame times | 4 of 10 starts with a ran without a core line, 7 lines in all (4, 1, 1, 1), each 2000.1-2000.9 ms; largest longest core wait in the 5 s vcores lines per start 515.0-2000.9 ms (median 1384.6); no starved grants; start to stable play 40.6-45.1 s (median 43.2; an ordinary run, no speed claim); machine load 7-44 at the starts; exit 0, no fatal line, kernel lines 0 in all 10 |
| 2026-11-14 | main 26a2eab003eeb18db69bcc41e1346dc4aec64b35 plus this change (the host/hos_thread.c change of the vcore starvation pull request, built alone on that commit: source stamp 6c3fbf7a, longshot sha256 2230b13d…), HOS_VCORE_STARVE_US at its default; tester mooki; run identifier ab1-fix, starts 01-10 | Arch Linux 7.2.15-1-omarchy, windowless/offscreen | RTX 4090 24 GB, NVIDIA 615.78.08 | Core i9-13900K, 96 GB RAM | 1920x1080, Custom preset, render 100%, game upscaler and frame generation Off, uncapped, real-time speed, HOS_FPS_MAX=1000; shaders prepared | the starting forest (child), a copied test save loaded from the title to play, 10 starts interleaved with run02 (the same build with HOS_VCORE_STARVE_US=0), three starts to one hold, 5 s standing after play began; counts and log lines, not frame times | 0 of 10 starts with a ran without a core line; largest longest core wait in the 5 s vcores lines per start 532.5-1002.0 ms (median 1000.4); 19 starved grants in 7 of the 10 starts; start to stable play 41.0-45.8 s (median 43.1; an ordinary run, no speed claim); machine load 7-49 at the starts; exit 0, no fatal line, kernel lines 0 in all 10 |
| 2026-11-14 | host and game folder built from main 9c9d69ed (v2.1.0-rc.3-463-g9c9d69ed); longshot_vk.so built alone from main 414f56e2 (nothing under gpu/ differs from 9c9d69ed) and from the three commits on it, 243c3a0d, same flags, each without a play profile and with the committed one (RX_VK_PGO, assets/profiles/longshot_vk.profdata.xz) |
SteamOS 3.9.5 (kernel 7.2.7-valve2), headless | Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel; the GPU’s clock fixed at 1600 MHz (Steam’s performance menu) | AMD Custom APU 0932 (4 cores, 8 threads), 16 GB shared with the GPU | Minimum preset (RX_VK_LOW_SPEC=1), handheld profile (HOS_HANDHELD=1), the game’s own upscaler, RX_VK_VRAM_MB=2400, RX_VK_TEX_MAX_DIM=1024, RX_VK_LOAD=create, RX_VK_READBACK_LATE=newest,3980, RX_VK_DESC_POOL_DIV=4, HOS_RESERVE_CORE=0; mesh programs on the compute path, no VRAM mirrors |
a save loaded by script, standing still: a heavy view (about 2,150 draws a frame) | p0lux run01, 720 presents measured a run, 2 runs each, main then the commits, the second round in the reverse order. Without a profile, main against the commits: 28.62 / 28.46 fps against 29.20 / 30.64; replay 31.12 / 31.17 ms against 30.04 / 28.94 ms; the render thread’s instructions a frame 62.38 / 61.99 M against 59.07 / 59.33 M, its cycles 79.15 / 78.77 M against 76.32 / 75.59 M. With the committed profile: 30.68 / 32.30 fps against 30.30 / 32.63; replay 26.99 / 26.50 ms against 26.49 / 26.02 ms; instructions 56.19 / 56.04 M against 54.61 / 54.44 M, cycles 68.99 / 68.76 M against 68.91 / 67.78 M. No errors in 9 starts |
| 2026-11-14 | host and game folder built from main 9c9d69ed (v2.1.0-rc.3-463-g9c9d69ed); longshot_vk.so built alone from main 9c9d69ed and from the staging arenas’ commit on it, d3004735, same flags |
SteamOS 3.9.5 (kernel 7.2.7-valve2), headless | Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel; the GPU’s clock fixed at 1600 MHz (Steam’s performance menu) | AMD Custom APU 0932 (4 cores, 8 threads), 16 GB shared with the GPU | Minimum preset (RX_VK_LOW_SPEC=1), handheld profile (HOS_HANDHELD=1), the game’s own upscaler, RX_VK_VRAM_MB=2400, RX_VK_TEX_MAX_DIM=1024, RX_VK_DESC_SLOTGEN=1, RX_VK_LOAD=create, HOS_RESERVE_CORE=0, RX_VK_FRAME_TRACE=1; RX_VK_STAGING_KEEP_MB unset (64 MB kept here) |
a save loaded by script, the arrival in the main outdoor area and in the starting forest: the whole run for the arenas, the 600 presents after the load’s transition frame for the frame times | p0lux run01, main then the commit, alternated with a third renderer (the all-zero check’s commit, its own file), 3 runs a scene. Staging arenas made in a run (the log’s a staging arena of 64 MB took lines, and their total): the main outdoor area 17 / 14 / 16 (242 / 202 / 225 ms) against 4 / 4 / 4 (58 / 45 / 65 ms); the starting forest 15 / 15 / 15 (184 / 198 / 192 ms) against 5 / 5 / 5 (78 / 72 / 64 ms). Upload time in the frames over 60 ms: the main outdoor area 341 / 262 / 292 ms against 163 / 170 / 171 ms; the starting forest 320 / 360 / 314 ms against 231 / 231 / 231 ms. Time in those frames: the main outdoor area 2.9 / 3.0 / 3.3 s against 2.9 / 3.4 / 3.0 s; the starting forest 6.0 / 9.1 / 8.6 s against 6.0 / 7.3 / 6.2 s, no difference beyond the spread between runs. drm-memory-gtt at the end of the run: the main outdoor area 3,290 / 3,249 / 3,317 MiB against 3,316 / 3,323 / 3,313 MiB; the starting forest 3,925 / 3,899 / 3,875 MiB against 3,926 / 3,866 / 3,925 MiB. No errors in 22 starts |
| 2026-11-14 | the change of f3264534 (a part ends only right before a plain draw) on 4da0099b: main cf605c85 with three changes not merged yet, the Battlemage float-add fix (3c8b0a99), the uniform snapshot on Arc discrete cards and XeSS (b46daa73) | Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | frame generation off; the default (RX_VK_SPLIT_BEFORE_DRAW on, as on every Intel GPU) |
the main outdoor area (child) and the Market (child) by the debug teleport, one launch, a recorded walk’s first 20 s replayed in each through the control bridge, a screenshot of every frame the bridge could take, each counted for white shards and the flagged ones looked at | the main outdoor area: 0 of 254 frames with white shards; the Market: 0 of 73; no device loss (run A4) |
| 2026-11-14 | the change of f3264534 (a part ends only right before a plain draw) on 4da0099b: main cf605c85 with three changes not merged yet, the Battlemage float-add fix (3c8b0a99), the uniform snapshot on Arc discrete cards and XeSS (b46daa73) | Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | frame generation off; RX_VK_SPLIT_BEFORE_DRAW=0 (main’s path) |
the main outdoor area (child) and the Market (child) by the debug teleport, one launch, a recorded walk’s first 20 s replayed in each through the control bridge, a screenshot of every frame the bridge could take, each counted for white shards and the flagged ones looked at | the main outdoor area: 8 of 264 frames with white shards; the Market: 0 of 211; no device loss (run A3) |
| 2026-11-14 | the change of f3264534 (a part ends only right before a plain draw) on 4da0099b: main cf605c85 with three changes not merged yet, the Battlemage float-add fix (3c8b0a99), the uniform snapshot on Arc discrete cards and XeSS (b46daa73) | Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | frame generation off; the default (RX_VK_SPLIT_BEFORE_DRAW on, as on every Intel GPU) |
the main outdoor area (child) and the Market (child) by the debug teleport, one launch, a recorded walk’s first 20 s replayed in each through the control bridge, a screenshot of every frame the bridge could take, each counted for white shards and the flagged ones looked at | the main outdoor area: 0 of 242 frames with white shards; the Market: 0 of 83; no device loss (run A2) |
| 2026-11-14 | the change of f3264534 (a part ends only right before a plain draw) on 4da0099b: main cf605c85 with three changes not merged yet, the Battlemage float-add fix (3c8b0a99), the uniform snapshot on Arc discrete cards and XeSS (b46daa73) | Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | frame generation off; RX_VK_SPLIT_BEFORE_DRAW=0 (main’s path) |
the main outdoor area (child) and the Market (child) by the debug teleport, one launch, a recorded walk’s first 20 s replayed in each through the control bridge, a screenshot of every frame the bridge could take, each counted for white shards and the flagged ones looked at | the main outdoor area: 4 of 232 frames with white shards; the Market: 1 of 25 (a broken frame right after the arrival), the rest of its walk a load; no device loss (run A1) |
| 2026-11-14 | the change of f3264534 (a part ends only right before a plain draw) on 4da0099b: main cf605c85 with three changes not merged yet, the Battlemage float-add fix (3c8b0a99), the uniform snapshot on Arc discrete cards and XeSS (b46daa73) | Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | frame generation off; the default (RX_VK_SPLIT_BEFORE_DRAW on, as on every Intel GPU) |
the starting forest (child), a recorded 30 s walk from the player character’s balcony replayed through the control bridge; no screenshots; fps from the log’s 5 s windows | 37.1, 51.6, 57.6, 58.2, 48.2, 47.6 fps, median 49.9 (run V10) |
| 2026-11-14 | the change of f3264534 (a part ends only right before a plain draw) on 4da0099b: main cf605c85 with three changes not merged yet, the Battlemage float-add fix (3c8b0a99), the uniform snapshot on Arc discrete cards and XeSS (b46daa73) | Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | frame generation off; RX_VK_SPLIT_BEFORE_DRAW=0 (main’s path) |
the starting forest (child), a recorded 30 s walk from the player character’s balcony replayed through the control bridge; no screenshots; fps from the log’s 5 s windows | 38.2, 46.6, 53.0, 53.7, 48.0, 48.3 fps, median 48.2 (run V9) |
| 2026-11-14 | the change of f3264534 (a part ends only right before a plain draw) on 4da0099b: main cf605c85 with three changes not merged yet, the Battlemage float-add fix (3c8b0a99), the uniform snapshot on Arc discrete cards and XeSS (b46daa73) | Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | frame generation off; the default (RX_VK_SPLIT_BEFORE_DRAW on, as on every Intel GPU) |
the starting forest (child), a recorded 30 s walk from the player character’s balcony replayed through the control bridge; no screenshots; fps from the log’s 5 s windows | 38.0, 52.7, 56.4, 56.6, 44.6, 57.9 fps, median 54.6 (run V8) |
| 2026-11-14 | the change of f3264534 (a part ends only right before a plain draw) on 4da0099b: main cf605c85 with three changes not merged yet, the Battlemage float-add fix (3c8b0a99), the uniform snapshot on Arc discrete cards and XeSS (b46daa73) | Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | frame generation off; RX_VK_SPLIT_BEFORE_DRAW=0 (main’s path) |
the starting forest (child), a recorded 30 s walk from the player character’s balcony replayed through the control bridge; no screenshots; fps from the log’s 5 s windows | 39.1, 50.6, 56.5, 56.1, 45.4, 52.7 fps, median 51.7 (run V7) |
| 2026-11-14 | the change of f3264534 (a part ends only right before a plain draw) on 4da0099b: main cf605c85 with three changes not merged yet, the Battlemage float-add fix (3c8b0a99), the uniform snapshot on Arc discrete cards and XeSS (b46daa73) | Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | frame generation off; the default (RX_VK_SPLIT_BEFORE_DRAW on, as on every Intel GPU) |
the starting forest (child), a recorded 30 s walk from the player character’s balcony replayed through the control bridge; a screenshot of every frame the control bridge could take, each counted for white shards | 0 of 382 frames with white shards (run V6) |
| 2026-11-14 | the change of f3264534 (a part ends only right before a plain draw) on 4da0099b: main cf605c85 with three changes not merged yet, the Battlemage float-add fix (3c8b0a99), the uniform snapshot on Arc discrete cards and XeSS (b46daa73) | Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | frame generation off; RX_VK_SPLIT_BEFORE_DRAW=0 (main’s path) |
the starting forest (child), a recorded 30 s walk from the player character’s balcony replayed through the control bridge; a screenshot of every frame the control bridge could take, each counted for white shards and the flagged ones looked at | 9 of 378 frames with white shards (run V5) |
| 2026-11-14 | the change of f3264534 (a part ends only right before a plain draw) on 4da0099b: main cf605c85 with three changes not merged yet, the Battlemage float-add fix (3c8b0a99), the uniform snapshot on Arc discrete cards and XeSS (b46daa73) | Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | frame generation off; the default (RX_VK_SPLIT_BEFORE_DRAW on, as on every Intel GPU) |
the starting forest (child), a recorded 30 s walk from the player character’s balcony replayed through the control bridge; a screenshot of every frame the control bridge could take, each counted for white shards | 0 of 387 frames with white shards (run V4) |
| 2026-11-14 | the change of f3264534 (a part ends only right before a plain draw) on 4da0099b: main cf605c85 with three changes not merged yet, the Battlemage float-add fix (3c8b0a99), the uniform snapshot on Arc discrete cards and XeSS (b46daa73) | Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | frame generation off; RX_VK_SPLIT_BEFORE_DRAW=0 (main’s path) |
the starting forest (child), a recorded 30 s walk from the player character’s balcony replayed through the control bridge; a screenshot of every frame the control bridge could take, each counted for white shards and the flagged ones looked at | 6 of 390 frames with white shards (run V3) |
| 2026-11-14 | the change of f3264534 (a part ends only right before a plain draw) on 4da0099b: main cf605c85 with three changes not merged yet, the Battlemage float-add fix (3c8b0a99), the uniform snapshot on Arc discrete cards and XeSS (b46daa73) | Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | frame generation off; the default (RX_VK_SPLIT_BEFORE_DRAW on, as on every Intel GPU) |
the starting forest (child), a recorded 30 s walk from the player character’s balcony replayed through the control bridge; a screenshot of every frame the control bridge could take, each counted for white shards | 0 of 332 frames with white shards (run V2) |
| 2026-11-14 | the change of f3264534 (a part ends only right before a plain draw) on 4da0099b: main cf605c85 with three changes not merged yet, the Battlemage float-add fix (3c8b0a99), the uniform snapshot on Arc discrete cards and XeSS (b46daa73) | Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | frame generation off; RX_VK_SPLIT_BEFORE_DRAW=0 (main’s path) |
the starting forest (child), a recorded 30 s walk from the player character’s balcony replayed through the control bridge; a screenshot of every frame the control bridge could take, each counted for white shards and the flagged ones looked at | 25 of 353 frames with white shards (run V1) |
| 2026-11-14 | fix/song-inventory-flags on main 5e072fc8c2adadba5ea610148be1c47800ba0cca with the local song fix; longshot v3.0.0-rc.3-64-g5e072fc8-dirty, source dd8ba6a9; built with dev_packs.py | CachyOS Linux, kernel 7.3.4-1-cachyos; Gamescope headless compositor with an X11 game window, separate from the locked Wayland desktop | NVIDIA GeForce RTX 5090, 620.70.06 | AMD Ryzen 9 9950X3D, 31 GiB RAM | 1600x900 window, 1600x896 output; DLSS Transformer 2 (M), 83% render resolution; frame generation Off; Debug tools On; RX_VK_PRECOMPILE=0; HOS_AUDIO=0 | the main outdoor area, the player character on horseback; Codex unattended run02 on Gustavo’s PC, using a copied save; opened only Longshot’s F1 menu, Debug > Inventory > Quest | the horse’s Song started Unlocked. Enter relocked then unlocked it; clicking the action button repeated both directions. Screenshots show Locked with Unlock and Unlocked with Relock; four log lines confirm the corresponding 1-to-0 and 0-to-1 flag changes. The original save folders were mounted read-only for the test process. After exit, all 14 source files retained their SHA-256, size and modification time, and the shared-save symlinks were unchanged. No performance measurement. |
| 2026-11-14 | main 0caf06184b391a5259d3741bb006ef5826eb6417 with the local song-inventory fix; longshot v3.0.0-rc.3-62-g0caf0618-dirty, source 71a08907 | CachyOS Linux, kernel 7.3.4-1-cachyos, Wayland | NVIDIA GeForce RTX 5090, 620.70.06 | AMD Ryzen 9 9950X3D, 31 GiB RAM | 3840x2160 output; Debug tools On; manual inventory-editor input | the main outdoor area, the player character; Gustavo’s manual run01, reported by the player and checked against the session log | The player confirmed that the editor updates correctly. The log records three Song of Storms relocks (had 1, now 0) and two unlocks (had 0, now 1). Before the fix, granting the horse’s Song succeeded in the game but the editor remained Locked because its pouch quantity was zero. Only Song of Storms was repeatedly toggled in this run; all twelve song mappings are covered by game-free regression tests. No performance measurement. |
| 2026-11-14 | Skip title Off 5c175a4727d82f85e425e85400dca17fea0750fb; native Linux release/profile build; tester mooki | Arch Linux 7.2.15-1-omarchy, windowless/offscreen | RTX 4090 24 GB, NVIDIA 615.78.08 | Core i9-13900K, 96 GB RAM | 1280x720, render 100%, game upscaler, frame generation/FPS overlay Off, real-time speed, HOS_FPS_MAX=1000, Skip title Off; saved V-Sync 0 (Auto), no swapchain | the starting forest, copied save and fixed camera, stationary; both builds warmed; A1/B1/A2/B2 in one exclusive quiet hold; 3 s scene settling then 25 s sample | Window 2 B2; median 13.540422 ms, mean 13.815604 ms, p99 20.891918 ms, from game-present-return CPU intervals, not display latency. Other CPU 1.696 mean / 1.766 five-second peak cores; load 12.32 to 8.24. Valid quiet window: no cost distinguishable from repeat spread (main 0.8566%, candidate 1.3907%), not proof of zero cost. Game exit 124 at its intended bound; complete journal, no fatal/device/driver fault. Scene pictures inspected. https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/399 |
| 2026-11-14 | Skip title Off 5c175a4727d82f85e425e85400dca17fea0750fb; native Linux release/profile build; tester mooki | Arch Linux 7.2.15-1-omarchy, windowless/offscreen | RTX 4090 24 GB, NVIDIA 615.78.08 | Core i9-13900K, 96 GB RAM | 1280x720, render 100%, game upscaler, frame generation/FPS overlay Off, real-time speed, HOS_FPS_MAX=1000, Skip title Off; saved V-Sync 0 (Auto), no swapchain | the starting forest, copied save and fixed camera, stationary; both builds warmed; A1/B1/A2/B2 in one exclusive quiet hold; 3 s scene settling then 25 s sample | Window 2 B1; median 13.730049 ms, mean 14.100743 ms, p99 22.106411 ms, from game-present-return CPU intervals, not display latency. Other CPU 1.768 mean / 1.893 five-second peak cores; load 7.04 to 7.31. Valid quiet window: no cost distinguishable from repeat spread (main 0.8566%, candidate 1.3907%), not proof of zero cost. Game exit 124 at its intended bound; complete journal, no fatal/device/driver fault. Scene pictures inspected. https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/399 |
| 2026-11-14 | main 0caf06184b391a5259d3741bb006ef5826eb6417; native Linux release/profile build; tester mooki | Arch Linux 7.2.15-1-omarchy, windowless/offscreen | RTX 4090 24 GB, NVIDIA 615.78.08 | Core i9-13900K, 96 GB RAM | 1280x720, render 100%, game upscaler, frame generation/FPS overlay Off, real-time speed, HOS_FPS_MAX=1000, Skip title Off; saved V-Sync 0 (Auto), no swapchain | the starting forest, copied save and fixed camera, stationary; both builds warmed; A1/B1/A2/B2 in one exclusive quiet hold; 3 s scene settling then 25 s sample | Window 2 A2; median 13.641001 ms, mean 13.981753 ms, p99 21.407780 ms, from game-present-return CPU intervals, not display latency. Other CPU 1.836 mean / 1.911 five-second peak cores; load 7.31 to 12.32. Valid quiet window: no cost distinguishable from repeat spread (main 0.8566%, candidate 1.3907%), not proof of zero cost. Game exit 124 at its intended bound; complete journal, no fatal/device/driver fault. Scene pictures inspected. https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/399 |
| 2026-11-14 | main 0caf06184b391a5259d3741bb006ef5826eb6417; native Linux release/profile build; tester mooki | Arch Linux 7.2.15-1-omarchy, windowless/offscreen | RTX 4090 24 GB, NVIDIA 615.78.08 | Core i9-13900K, 96 GB RAM | 1280x720, render 100%, game upscaler, frame generation/FPS overlay Off, real-time speed, HOS_FPS_MAX=1000, Skip title Off; saved V-Sync 0 (Auto), no swapchain | the starting forest, copied save and fixed camera, stationary; both builds warmed; A1/B1/A2/B2 in one exclusive quiet hold; 3 s scene settling then 25 s sample | Window 2 A1; median 13.758351 ms, mean 14.152099 ms, p99 21.785438 ms, from game-present-return CPU intervals, not display latency. Other CPU 1.767 mean / 1.904 five-second peak cores; load 1.26 to 7.04. Valid quiet window: no cost distinguishable from repeat spread (main 0.8566%, candidate 1.3907%), not proof of zero cost. Game exit 124 at its intended bound; complete journal, no fatal/device/driver fault. Scene pictures inspected. https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/399 |
| 2026-11-14 | Skip title Off 5c175a4727d82f85e425e85400dca17fea0750fb; native Linux release/profile build; tester mooki | Arch Linux 7.2.15-1-omarchy, windowless/offscreen | RTX 4090 24 GB, NVIDIA 615.78.08 | Core i9-13900K, 96 GB RAM | 1280x720, render 100%, game upscaler, frame generation/FPS overlay Off, real-time speed, HOS_FPS_MAX=1000, Skip title Off; saved V-Sync 0 (Auto), no swapchain | the starting forest, copied save and fixed camera, stationary; both builds warmed; A1/B1/A2/B2 in one exclusive quiet hold; 3 s scene settling then 25 s sample | Window 1 B2; median 13.653566 ms, mean 14.017513 ms, p99 21.309892 ms, from game-present-return CPU intervals, not display latency. Other CPU 1.788 mean / 1.868 five-second peak cores; load 17.73 to 9.92. REJECTED WINDOW: A2 was disturbed (7.942 mean / 28.797 peak other-CPU cores), so none of this window establishes cost. Game exit 124 at its intended bound; complete journal, no fatal/device/driver fault. Scene pictures inspected. https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/399 |
| 2026-11-14 | Skip title Off 5c175a4727d82f85e425e85400dca17fea0750fb; native Linux release/profile build; tester mooki | Arch Linux 7.2.15-1-omarchy, windowless/offscreen | RTX 4090 24 GB, NVIDIA 615.78.08 | Core i9-13900K, 96 GB RAM | 1280x720, render 100%, game upscaler, frame generation/FPS overlay Off, real-time speed, HOS_FPS_MAX=1000, Skip title Off; saved V-Sync 0 (Auto), no swapchain | the starting forest, copied save and fixed camera, stationary; both builds warmed; A1/B1/A2/B2 in one exclusive quiet hold; 3 s scene settling then 25 s sample | Window 1 B1; median 13.463051 ms, mean 13.837404 ms, p99 21.195355 ms, from game-present-return CPU intervals, not display latency. Other CPU 1.675 mean / 1.776 five-second peak cores; load 5.81 to 6.90. REJECTED WINDOW: A2 was disturbed (7.942 mean / 28.797 peak other-CPU cores), so none of this window establishes cost. Game exit 124 at its intended bound; complete journal, no fatal/device/driver fault. Scene pictures inspected. https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/399 |
| 2026-11-14 | main 0caf06184b391a5259d3741bb006ef5826eb6417; native Linux release/profile build; tester mooki | Arch Linux 7.2.15-1-omarchy, windowless/offscreen | RTX 4090 24 GB, NVIDIA 615.78.08 | Core i9-13900K, 96 GB RAM | 1280x720, render 100%, game upscaler, frame generation/FPS overlay Off, real-time speed, HOS_FPS_MAX=1000, Skip title Off; saved V-Sync 0 (Auto), no swapchain | the starting forest, copied save and fixed camera, stationary; both builds warmed; A1/B1/A2/B2 in one exclusive quiet hold; 3 s scene settling then 25 s sample | Window 1 A2; median 13.905075 ms, mean 17.108690 ms, p99 102.825698 ms, from game-present-return CPU intervals, not display latency. Other CPU 7.942 mean / 28.797 five-second peak cores; load 6.90 to 17.73. REJECTED WINDOW: A2 was disturbed (7.942 mean / 28.797 peak other-CPU cores), so none of this window establishes cost. Game exit 124 at its intended bound; complete journal, no fatal/device/driver fault. Scene pictures inspected. https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/399 |
| 2026-11-14 | main 0caf06184b391a5259d3741bb006ef5826eb6417; native Linux release/profile build; tester mooki | Arch Linux 7.2.15-1-omarchy, windowless/offscreen | RTX 4090 24 GB, NVIDIA 615.78.08 | Core i9-13900K, 96 GB RAM | 1280x720, render 100%, game upscaler, frame generation/FPS overlay Off, real-time speed, HOS_FPS_MAX=1000, Skip title Off; saved V-Sync 0 (Auto), no swapchain | the starting forest, copied save and fixed camera, stationary; both builds warmed; A1/B1/A2/B2 in one exclusive quiet hold; 3 s scene settling then 25 s sample | Window 1 A1; median 13.618077 ms, mean 13.881730 ms, p99 21.311780 ms, from game-present-return CPU intervals, not display latency. Other CPU 1.689 mean / 1.746 five-second peak cores; load 1.47 to 5.81. REJECTED WINDOW: A2 was disturbed (7.942 mean / 28.797 peak other-CPU cores), so none of this window establishes cost. Game exit 124 at its intended bound; complete journal, no fatal/device/driver fault. Scene pictures inspected. https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/399 |
| 2026-11-14 | development base 514591a3 + psycho-scene-variants (matcher fix; measured header b20f94fdf355a01d) | Windows, build 26300, 26H2; windowed | NVIDIA RTX 4070 / 621.01 | Intel Core i7-13700F / 32 GB | Custom; 2560x1440, 16:9; Native HDR 200/1000 nits; PsychoV35; DLAA K; FG Off; game cap 61; RX_VK_PSYCHOV35_DLSS=1 | Lon Lon Ranch daylight alley, fresh copied checkpoint, two-pitch yaw sweep; run b2 (diagv2-fixed03) | 49 stable-play shader switches; 0 large chroma steps (>1 capture code, median over 98 blocks within +/-5 frames); median absolute step 0.03251, maximum 0.08876; 4793 complete capture records; no FPS claim |
| 2026-11-14 | development base 514591a3 + psycho-scene-variants (matcher fix; measured header b20f94fdf355a01d) | Windows, build 26300, 26H2; windowed | NVIDIA RTX 4070 / 621.01 | Intel Core i7-13700F / 32 GB | Custom; 2560x1440, 16:9; Native HDR 200/1000 nits; PsychoV35; DLAA K; FG Off; game cap 61; RX_VK_PSYCHOV35_DLSS=1 | Lon Lon Ranch daylight alley, fresh copied checkpoint, two-pitch yaw sweep; run b1 (diagv2-fixed02) | 31 stable-play shader switches; 0 large chroma steps (>1 capture code, median over 98 blocks within +/-5 frames); median absolute step 0.03625, maximum 0.12250; 5129 complete capture records; no FPS claim |
| 2026-11-14 | development base 514591a3 + psycho-tone-diag-v2 (original matcher, identical to main 0caf0618) | Windows, build 26300, 26H2; windowed | NVIDIA RTX 4070 / 621.01 | Intel Core i7-13700F / 32 GB | Custom; 2560x1440, 16:9; Native HDR 200/1000 nits; PsychoV35; DLAA K; FG Off; game cap 61; RX_VK_PSYCHOV35_DLSS=1 | Lon Lon Ranch daylight alley, fresh copied checkpoint, two-pitch yaw sweep; run a2 (diag-original04) | 49 stable-play shader switches; 37 large chroma steps (>1 capture code, median over 98 blocks within +/-5 frames); median absolute step 3.05500, maximum 4.37625; 5059 complete capture records; no FPS claim |
| 2026-11-14 | development base 514591a3 + psycho-tone-diag-v2 (original matcher, identical to main 0caf0618) | Windows, build 26300, 26H2; windowed | NVIDIA RTX 4070 / 621.01 | Intel Core i7-13700F / 32 GB | Custom; 2560x1440, 16:9; Native HDR 200/1000 nits; PsychoV35; DLAA K; FG Off; game cap 61; RX_VK_PSYCHOV35_DLSS=1 | Lon Lon Ranch daylight alley, fresh copied checkpoint, two-pitch yaw sweep; run a1 (diagv2-hunt03) | 45 stable-play shader switches; 45 large chroma steps (>1 capture code, median over 98 blocks within +/-5 frames); median absolute step 2.12625, maximum 5.06750; 5279 complete capture records; no FPS claim |
| 2026-11-14 | development base 514591a3 + PR400 matcher and identical private diagnostics + TAAU fix e1da8a66 | Windows kernel 26300 / 26H2, windowed | NVIDIA RTX 4070 / 621.01 | Intel Core i7-13700F / 32 GB | 2560x1440 output, 2560x1421 client; Native HDR 200/1000 nits; game upscaler; PsychoV35, Vanilla control; 75% and 100%; FG Off, cap 61; RX_VK_TAAU_KEYS=fd029a317a9a4b94 to exercise replacement on NVIDIA; compatibility flag 0 | Lon Lon Ranch daylight alley, fresh copied checkpoint; b2 in original/fixed/original/fixed order | Run b2: 75% PsychoV35 ROI exact-black fraction 0.000000; RGBA16F helper logged, scene visible; Vanilla 75% control and Psycho/Vanilla/Psycho and 75/100/75 switches captured; five 10-second settled phases; normal exit, no device-lost/pipeline-failure line; no FPS claim. |
| 2026-11-14 | development base 514591a3 + PR400 matcher and identical private diagnostics + TAAU fix e1da8a66 | Windows kernel 26300 / 26H2, windowed | NVIDIA RTX 4070 / 621.01 | Intel Core i7-13700F / 32 GB | 2560x1440 output, 2560x1421 client; Native HDR 200/1000 nits; game upscaler; PsychoV35, Vanilla control; 75% and 100%; FG Off, cap 61; RX_VK_TAAU_KEYS=fd029a317a9a4b94 to exercise replacement on NVIDIA; compatibility flag 0 | Lon Lon Ranch daylight alley, fresh copied checkpoint; b1 in original/fixed/original/fixed order | Run b1: 75% PsychoV35 ROI exact-black fraction 0.000000; RGBA16F helper logged, scene visible; Vanilla 75% control and Psycho/Vanilla/Psycho and 75/100/75 switches captured; five 10-second settled phases; normal exit, no device-lost/pipeline-failure line; no FPS claim. |
| 2026-11-14 | development base 514591a3 + PR400 matcher and identical private diagnostics; original TAAU replacement | Windows kernel 26300 / 26H2, windowed | NVIDIA RTX 4070 / 621.01 | Intel Core i7-13700F / 32 GB | 2560x1440 output, 2560x1421 client; Native HDR 200/1000 nits; game upscaler; PsychoV35, Vanilla control; 75% and 100%; FG Off, cap 61; RX_VK_TAAU_KEYS=fd029a317a9a4b94 to exercise replacement on NVIDIA; compatibility flag 0 | Lon Lon Ranch daylight alley, fresh copied checkpoint; a2 in original/fixed/original/fixed order | Run a2: 75% PsychoV35 ROI exact-black fraction 1.000000; no RGBA16F helper, black scene with HUD; Vanilla 75% control and Psycho/Vanilla/Psycho and 75/100/75 switches captured; five 10-second settled phases; normal exit, no device-lost/pipeline-failure line; no FPS claim. |
| 2026-11-14 | development base 514591a3 + PR400 matcher and identical private diagnostics; original TAAU replacement | Windows kernel 26300 / 26H2, windowed | NVIDIA RTX 4070 / 621.01 | Intel Core i7-13700F / 32 GB | 2560x1440 output, 2560x1421 client; Native HDR 200/1000 nits; game upscaler; PsychoV35, Vanilla control; 75% and 100%; FG Off, cap 61; RX_VK_TAAU_KEYS=fd029a317a9a4b94 to exercise replacement on NVIDIA; compatibility flag 0 | Lon Lon Ranch daylight alley, fresh copied checkpoint; a1 in original/fixed/original/fixed order | Run a1: 75% PsychoV35 ROI exact-black fraction 1.000000; no RGBA16F helper, black scene with HUD; Vanilla 75% control and Psycho/Vanilla/Psycho and 75/100/75 switches captured; five 10-second settled phases; normal exit, no device-lost/pipeline-failure line; no FPS claim. |
| 2026-11-14 | development base 514591a3 + matcher fix from PR400 and identical private route logging + Off fix 5b9c7666 | Windows build 26300 / 26H2, windowed | NVIDIA RTX 4070 / 621.01 | Intel Core i7-13700F / 32 GB | 2560x1440 output, 2560x1421 client; Native HDR 200/1000 nits; Off; PsychoV35 then Vanilla and back; 100%, 75%, 67%; FG Off, cap 61; RX_VK_PSYCHOV35_DLSS=0 | Lon Lon Ranch daylight alley, fresh copied checkpoint; b2 in interleaved original/fixed/original/fixed order | Run b2: RGBA16F Off bypass confirmed at 3/3 scales (100/75/67%); Vanilla 100% bypass works; tone mapper switched Vanilla and back; five 10-second settled phases captured, normal exit, no device-lost or shader-failure line. No FPS claim. |
| 2026-11-14 | development base 514591a3 + matcher fix from PR400 and identical private route logging + Off fix 5b9c7666 | Windows build 26300 / 26H2, windowed | NVIDIA RTX 4070 / 621.01 | Intel Core i7-13700F / 32 GB | 2560x1440 output, 2560x1421 client; Native HDR 200/1000 nits; Off; PsychoV35 then Vanilla and back; 100%, 75%, 67%; FG Off, cap 61; RX_VK_PSYCHOV35_DLSS=0 | Lon Lon Ranch daylight alley, fresh copied checkpoint; b1 in interleaved original/fixed/original/fixed order | Run b1: RGBA16F Off bypass confirmed at 3/3 scales (100/75/67%); Vanilla 100% bypass works; tone mapper switched Vanilla and back; five 10-second settled phases captured, normal exit, no device-lost or shader-failure line. No FPS claim. |
| 2026-11-14 | development base 514591a3 + matcher fix from PR400 and identical private route logging; original Off | Windows build 26300 / 26H2, windowed | NVIDIA RTX 4070 / 621.01 | Intel Core i7-13700F / 32 GB | 2560x1440 output, 2560x1421 client; Native HDR 200/1000 nits; Off; PsychoV35 then Vanilla and back; 100%, 75%, 67%; FG Off, cap 61; RX_VK_PSYCHOV35_DLSS=0 | Lon Lon Ranch daylight alley, fresh copied checkpoint; a2 in interleaved original/fixed/original/fixed order | Run a2: RGBA16F Off bypass confirmed at 0/3 scales (100/75/67%); Vanilla 100% bypass works; tone mapper switched Vanilla and back; five 10-second settled phases captured, normal exit, no device-lost or shader-failure line. No FPS claim. |
| 2026-11-14 | development base 514591a3 + matcher fix from PR400 and identical private route logging; original Off | Windows build 26300 / 26H2, windowed | NVIDIA RTX 4070 / 621.01 | Intel Core i7-13700F / 32 GB | 2560x1440 output, 2560x1421 client; Native HDR 200/1000 nits; Off; PsychoV35 then Vanilla and back; 100%, 75%, 67%; FG Off, cap 61; RX_VK_PSYCHOV35_DLSS=0 | Lon Lon Ranch daylight alley, fresh copied checkpoint; a1 in interleaved original/fixed/original/fixed order | Run a1: RGBA16F Off bypass confirmed at 0/3 scales (100/75/67%); Vanilla 100% bypass works; tone mapper switched Vanilla and back; five 10-second settled phases captured, normal exit, no device-lost or shader-failure line. No FPS claim. |
| 2026-11-14 | development base514591a3 + PR400 + FSR3 upscaling e7efc7c8 + FG fix f072407f | Windows kernel26300 /26H2 | RTX4070 / NVIDIA621.01 | Intel Core i7-13700F /32GB | 2560x1440; Native HDR200/1000 nits; PsychoV35; FSR3.1 at67%; FG2x, Built-in provider, NVIDIA FG disabled; both compatibility/FSR FG opt-ins1; inline made capture every120 presents | Lon Lon Ranch alley, copied checkpoint; capture-b; 12s settle, right stick +0.25 for120 frames then -0.25 for120 frames,12s return hold | 23 made PNGs; last capture interval 595/600 AMD-generated, 0 missing inputs. Capture/readback adds hitches/reset fallbacks; these counts are separate from normal-path paired measurements. Selected stationary/moving generated frames inspected: visible scene, no all-black/obvious global colour corruption; not calibrated HDR or temporal-quality proof. Normal exit. |
| 2026-11-14 | development base514591a3 + PR400 + FSR3 upscaling e7efc7c8; original FG | Windows kernel26300 /26H2 | RTX4070 / NVIDIA621.01 | Intel Core i7-13700F /32GB | 2560x1440; Native HDR200/1000 nits; PsychoV35; FSR3.1 at67%; FG2x, Built-in provider, NVIDIA FG disabled; both compatibility/FSR FG opt-ins1; inline made capture every120 presents | Lon Lon Ranch alley, copied checkpoint; capture-a; 12s settle, right stick +0.25 for120 frames then -0.25 for120 frames,12s return hold | 23 made PNGs; last capture interval 0/600 AMD-generated, 600 missing inputs. Capture/readback adds hitches/reset fallbacks; these counts are separate from normal-path paired measurements. Selected stationary/moving generated frames inspected: visible scene, no all-black/obvious global colour corruption; not calibrated HDR or temporal-quality proof. Normal exit. |
| 2026-11-14 | development base514591a3 + PR400 matcher + FSR3 upscaling fix e7efc7c8 and identical private diagnostics + FG fix f072407f | Windows kernel26300 / 26H2, windowed | NVIDIA RTX4070 / 621.01 | Intel Core i7-13700F /32GB | 2560x1440 output,2560x1421 client; Native HDR200/1000 nits (800 at83%), Vanilla SDR control; FSR3.1 upscaling,67/83/100%; FSR FG2x and Off/on; provider Built-in, NVIDIA FG disabled; sharpness0, cap61; RX_VK_FSR_FG=1 and RX_VK_PSYCHOV35_DLSS=1 | Lon Lon Ranch daylight alley, fresh copied checkpoint; runb4 in original/fixed/original/fixed order; eight phases advancing at least1250 presents each | psycho67 600/600 AMD-made, 0 missing-input fallbacks; psycho83 600/600 AMD-made, 0 missing-input fallbacks; psycho100 600/600 AMD-made, 0 missing-input fallbacks; vanilla67 600/600 AMD-made, 0 missing-input fallbacks; sdr67 600/600 AMD-made, 0 missing-input fallbacks; psycho-return67 600/600 AMD-made, 0 missing-input fallbacks; fg-off Off, no AMD generation interval; fg-return 600/600 AMD-made, 0 missing-input fallbacks; PQ/sRGB metadata and1000/800-nit changes logged; normal exit, no device-loss/pipeline-error line; capped functional check, no FPS or calibrated HDR claim |
| 2026-11-14 | development base514591a3 + PR400 matcher + FSR3 upscaling fix e7efc7c8 and identical private diagnostics + FG fix f072407f | Windows kernel26300 / 26H2, windowed | NVIDIA RTX4070 / 621.01 | Intel Core i7-13700F /32GB | 2560x1440 output,2560x1421 client; Native HDR200/1000 nits (800 at83%), Vanilla SDR control; FSR3.1 upscaling,67/83/100%; FSR FG2x and Off/on; provider Built-in, NVIDIA FG disabled; sharpness0, cap61; RX_VK_FSR_FG=1 and RX_VK_PSYCHOV35_DLSS=1 | Lon Lon Ranch daylight alley, fresh copied checkpoint; runb3 in original/fixed/original/fixed order; eight phases advancing at least1250 presents each | psycho67 600/600 AMD-made, 0 missing-input fallbacks; psycho83 600/600 AMD-made, 0 missing-input fallbacks; psycho100 600/600 AMD-made, 0 missing-input fallbacks; vanilla67 600/600 AMD-made, 0 missing-input fallbacks; sdr67 600/600 AMD-made, 0 missing-input fallbacks; psycho-return67 600/600 AMD-made, 0 missing-input fallbacks; fg-off Off, no AMD generation interval; fg-return 600/600 AMD-made, 0 missing-input fallbacks; PQ/sRGB metadata and1000/800-nit changes logged; normal exit, no device-loss/pipeline-error line; capped functional check, no FPS or calibrated HDR claim |
| 2026-11-14 | development base514591a3 + PR400 matcher + FSR3 upscaling fix e7efc7c8 and identical private diagnostics; original FSR FG | Windows kernel26300 / 26H2, windowed | NVIDIA RTX4070 / 621.01 | Intel Core i7-13700F /32GB | 2560x1440 output,2560x1421 client; Native HDR200/1000 nits (800 at83%), Vanilla SDR control; FSR3.1 upscaling,67/83/100%; FSR FG2x and Off/on; provider Built-in, NVIDIA FG disabled; sharpness0, cap61; RX_VK_FSR_FG=1 and RX_VK_PSYCHOV35_DLSS=1 | Lon Lon Ranch daylight alley, fresh copied checkpoint; runa4 in original/fixed/original/fixed order; eight phases advancing at least1250 presents each | psycho67 0/600 AMD-made, 600 missing-input fallbacks; psycho83 600/600 AMD-made, 0 missing-input fallbacks; psycho100 600/600 AMD-made, 0 missing-input fallbacks; vanilla67 600/600 AMD-made, 0 missing-input fallbacks; sdr67 600/600 AMD-made, 0 missing-input fallbacks; psycho-return67 0/600 AMD-made, 600 missing-input fallbacks; fg-off Off, no AMD generation interval; fg-return 0/600 AMD-made, 600 missing-input fallbacks; Original HDR constant blocks remain sRGB despite native PQ presentation; normal exit, no device-loss/pipeline-error line; capped functional check, no FPS or calibrated HDR claim |
| 2026-11-14 | development base514591a3 + PR400 matcher + FSR3 upscaling fix e7efc7c8 and identical private diagnostics; original FSR FG | Windows kernel26300 / 26H2, windowed | NVIDIA RTX4070 / 621.01 | Intel Core i7-13700F /32GB | 2560x1440 output,2560x1421 client; Native HDR200/1000 nits (800 at83%), Vanilla SDR control; FSR3.1 upscaling,67/83/100%; FSR FG2x and Off/on; provider Built-in, NVIDIA FG disabled; sharpness0, cap61; RX_VK_FSR_FG=1 and RX_VK_PSYCHOV35_DLSS=1 | Lon Lon Ranch daylight alley, fresh copied checkpoint; runa3 in original/fixed/original/fixed order; eight phases advancing at least1250 presents each | psycho67 0/600 AMD-made, 600 missing-input fallbacks; psycho83 600/600 AMD-made, 0 missing-input fallbacks; psycho100 600/600 AMD-made, 0 missing-input fallbacks; vanilla67 600/600 AMD-made, 0 missing-input fallbacks; sdr67 600/600 AMD-made, 0 missing-input fallbacks; psycho-return67 0/600 AMD-made, 600 missing-input fallbacks; fg-off Off, no AMD generation interval; fg-return 0/600 AMD-made, 600 missing-input fallbacks; Original HDR constant blocks remain sRGB despite native PQ presentation; normal exit, no device-loss/pipeline-error line; capped functional check, no FPS or calibrated HDR claim |
| 2026-11-14 | development base 514591a3 + PR400 matcher and identical private diagnostics + FSR 3.1 fix e7efc7c8 | Windows kernel 26300 / 26H2, windowed | NVIDIA RTX 4070 / 621.01 | Intel Core i7-13700F / 32 GB | 2560x1440 output, 2560x1421 client; Native HDR 200/1000 nits; FSR 3.1 upscaling; PsychoV35 67/83/100%, Vanilla 67%; FG Off, cap 61, sharpness 0; RX_VK_PSYCHOV35_DLSS=1 | Lon Lon Ranch daylight alley, fresh copied checkpoint; b2 in original/fixed/original/fixed order | Run b2: last complete 600-frame status interval per phase: psycho67 600/600, psycho83 600/600, psycho100 600/600, vanilla67 600/600, psycho-return67 600/600; RGBA16F linear route logged; each phase advanced at least 1250 presents; normal exit, no device-lost/pipeline-failure line; no FPS claim. |
| 2026-11-14 | development base 514591a3 + PR400 matcher and identical private diagnostics + FSR 3.1 fix e7efc7c8 | Windows kernel 26300 / 26H2, windowed | NVIDIA RTX 4070 / 621.01 | Intel Core i7-13700F / 32 GB | 2560x1440 output, 2560x1421 client; Native HDR 200/1000 nits; FSR 3.1 upscaling; PsychoV35 67/83/100%, Vanilla 67%; FG Off, cap 61, sharpness 0; RX_VK_PSYCHOV35_DLSS=1 | Lon Lon Ranch daylight alley, fresh copied checkpoint; b1 in original/fixed/original/fixed order | Run b1: last complete 600-frame status interval per phase: psycho67 600/600, psycho83 600/600, psycho100 600/600, vanilla67 600/600, psycho-return67 600/600; RGBA16F linear route logged; each phase advanced at least 1250 presents; normal exit, no device-lost/pipeline-failure line; no FPS claim. |
| 2026-11-14 | development base 514591a3 + PR400 matcher and identical private diagnostics; original FSR 3.1 | Windows kernel 26300 / 26H2, windowed | NVIDIA RTX 4070 / 621.01 | Intel Core i7-13700F / 32 GB | 2560x1440 output, 2560x1421 client; Native HDR 200/1000 nits; FSR 3.1 upscaling; PsychoV35 67/83/100%, Vanilla 67%; FG Off, cap 61, sharpness 0; RX_VK_PSYCHOV35_DLSS=1 | Lon Lon Ranch daylight alley, fresh copied checkpoint; a2 in original/fixed/original/fixed order | Run a2: last complete 600-frame status interval per phase: psycho67 0/600, psycho83 0/600, psycho100 0/600, vanilla67 600/600, psycho-return67 0/600; no scene camera in PsychoV35; game network fallback; each phase advanced at least 1250 presents; normal exit, no device-lost/pipeline-failure line; no FPS claim. |
| 2026-11-14 | development base 514591a3 + PR400 matcher and identical private diagnostics; original FSR 3.1 | Windows kernel 26300 / 26H2, windowed | NVIDIA RTX 4070 / 621.01 | Intel Core i7-13700F / 32 GB | 2560x1440 output, 2560x1421 client; Native HDR 200/1000 nits; FSR 3.1 upscaling; PsychoV35 67/83/100%, Vanilla 67%; FG Off, cap 61, sharpness 0; RX_VK_PSYCHOV35_DLSS=1 | Lon Lon Ranch daylight alley, fresh copied checkpoint; a1 in original/fixed/original/fixed order | Run a1: last complete 600-frame status interval per phase: psycho67 0/600, psycho83 0/600, psycho100 0/600, vanilla67 600/600, psycho-return67 0/600; no scene camera in PsychoV35; game network fallback; each phase advanced at least 1250 presents; normal exit, no device-lost/pipeline-failure line; no FPS claim. |
| 2026-11-14 | rollout3fa9b801, combined prerequisites757ba924; isolated renderer pack on unchanged development base514591a3 with common scene matcher, upscaler and FG fixes;19 relevant production functions/helpers match current source; base pipeline-creation wrappers retained | Windows kernel26300 /26H2, windowed2560x1421 client | RTX4070 12GB, NVIDIA621.01 | Core i7-13700F,32GB | 2560x1440 output; RX_VK_PSYCHOV35_DLSS=unset; FSR67% with AMD FG2x, DLSS67%, Off100%, forcedTAAU75%, Vanilla FSR67%, SDR FSR67%, returnPsycho FSR67% with AMD FG2x; FG Off in non-Psycho-FSR phases; paper white200/peak1000,sharpness0,cap61 | Daylight Lon Lon Ranch alley, fresh copied checkpoint per process; interleaved explicit1/unset pairs then rollback,NGX and cold-DLAA controls; FG phases1250+,other phases650+ presents | Run unset-d3: 7 phases recorded, all stable play; normal exit. AMD [made, requests, missing]: fsr67 [600, 600, 0]; return-fsr67 [600, 600, 0]; last GPU-time samples by phase (ms): fsr67: 10.6, dlss67: 9.1, off100: 9.2, taau75: 8.1, vanilla67: 8.7, sdr67: 8.6, return-fsr67: 10.9. Capped functional/parity run, not an uncapped performance or equivalence claim; no device loss. |
| 2026-11-14 | rollout3fa9b801, combined prerequisites757ba924; isolated renderer pack on unchanged development base514591a3 with common scene matcher, upscaler and FG fixes;19 relevant production functions/helpers match current source; base pipeline-creation wrappers retained | Windows kernel26300 /26H2, windowed2560x1421 client | RTX4070 12GB, NVIDIA621.01 | Core i7-13700F,32GB | 2560x1440 output; RX_VK_PSYCHOV35_DLSS=unset; FSR67% with AMD FG2x, DLSS67%, Off100%, forcedTAAU75%, Vanilla FSR67%, SDR FSR67%, returnPsycho FSR67% with AMD FG2x; FG Off in non-Psycho-FSR phases; paper white200/peak1000,sharpness0,cap61 | Daylight Lon Lon Ranch alley, fresh copied checkpoint per process; interleaved explicit1/unset pairs then rollback,NGX and cold-DLAA controls; FG phases1250+,other phases650+ presents | Run unset-b3: 7 phases recorded, all stable play; normal exit. AMD [made, requests, missing]: fsr67 [600, 600, 0]; return-fsr67 [600, 600, 0]; last GPU-time samples by phase (ms): fsr67: 10.7, dlss67: 9.0, off100: 9.1, taau75: 8.2, vanilla67: 8.6, sdr67: 8.7, return-fsr67: 10.8. Capped functional/parity run, not an uncapped performance or equivalence claim; no device loss. |
| 2026-11-14 | rollout3fa9b801, combined prerequisites757ba924; isolated renderer pack on unchanged development base514591a3 with common scene matcher, upscaler and FG fixes;19 relevant production functions/helpers match current source; base pipeline-creation wrappers retained | Windows kernel26300 /26H2, windowed2560x1421 client | RTX4070 12GB, NVIDIA621.01 | Core i7-13700F,32GB | 2560x1440 output; RX_VK_PSYCHOV35_DLSS=unset; DLSS67%, NGX FG2x, Auto provider, AMD FG disabled; Native HDR PsychoV35; paper white200/peak1000,sharpness0,cap61 | Daylight Lon Lon Ranch alley, fresh copied checkpoint per process; interleaved explicit1/unset pairs then rollback,NGX and cold-DLAA controls; FG phases1250+,other phases650+ presents | Run nvidia-unset-f3: 1 phases recorded, all stable play; normal exit. AMD [made, requests, missing]: not enabled; last GPU-time samples by phase (ms): dlss67: 12.0. Capped functional/parity run, not an uncapped performance or equivalence claim; no device loss. |
| 2026-11-14 | rollout3fa9b801, combined prerequisites757ba924; isolated renderer pack on unchanged development base514591a3 with common scene matcher, upscaler and FG fixes;19 relevant production functions/helpers match current source; base pipeline-creation wrappers retained | Windows kernel26300 /26H2, windowed2560x1421 client | RTX4070 12GB, NVIDIA621.01 | Core i7-13700F,32GB | 2560x1440 output; RX_VK_PSYCHOV35_DLSS=1; FSR67% with AMD FG2x, DLSS67%, Off100%, forcedTAAU75%, Vanilla FSR67%, SDR FSR67%, returnPsycho FSR67% with AMD FG2x; FG Off in non-Psycho-FSR phases; paper white200/peak1000,sharpness0,cap61 | Daylight Lon Lon Ranch alley, fresh copied checkpoint per process; interleaved explicit1/unset pairs then rollback,NGX and cold-DLAA controls; FG phases1250+,other phases650+ presents | Run explicit1-c3: 7 phases recorded, all stable play; normal exit. AMD [made, requests, missing]: fsr67 [600, 600, 0]; return-fsr67 [600, 600, 0]; last GPU-time samples by phase (ms): fsr67: 10.8, dlss67: 9.0, off100: 9.1, taau75: 8.1, vanilla67: 8.6, sdr67: 8.5, return-fsr67: 10.8. Capped functional/parity run, not an uncapped performance or equivalence claim; no device loss. |
| 2026-11-14 | rollout3fa9b801, combined prerequisites757ba924; isolated renderer pack on unchanged development base514591a3 with common scene matcher, upscaler and FG fixes;19 relevant production functions/helpers match current source; base pipeline-creation wrappers retained | Windows kernel26300 /26H2, windowed2560x1421 client | RTX4070 12GB, NVIDIA621.01 | Core i7-13700F,32GB | 2560x1440 output; RX_VK_PSYCHOV35_DLSS=1; FSR67% with AMD FG2x, DLSS67%, Off100%, forcedTAAU75%, Vanilla FSR67%, SDR FSR67%, returnPsycho FSR67% with AMD FG2x; FG Off in non-Psycho-FSR phases; paper white200/peak1000,sharpness0,cap61 | Daylight Lon Lon Ranch alley, fresh copied checkpoint per process; interleaved explicit1/unset pairs then rollback,NGX and cold-DLAA controls; FG phases1250+,other phases650+ presents | Run explicit1-a3: 7 phases recorded, all stable play; normal exit. AMD [made, requests, missing]: fsr67 [600, 600, 0]; return-fsr67 [600, 600, 0]; last GPU-time samples by phase (ms): fsr67: 10.9, dlss67: 9.0, off100: 9.5, taau75: 8.3, vanilla67: 8.7, sdr67: 8.6, return-fsr67: 10.9. Capped functional/parity run, not an uncapped performance or equivalence claim; no device loss. |
| 2026-11-14 | rollout3fa9b801, combined prerequisites757ba924; isolated renderer pack on unchanged development base514591a3 with common scene matcher, upscaler and FG fixes;19 relevant production functions/helpers match current source; base pipeline-creation wrappers retained | Windows kernel26300 /26H2, windowed2560x1421 client | RTX4070 12GB, NVIDIA621.01 | Core i7-13700F,32GB | 2560x1440 output; RX_VK_PSYCHOV35_DLSS=unset; DLAA from saved Native choice, FG Off; Native HDR PsychoV35; paper white200/peak1000,sharpness0,cap61 | Daylight Lon Lon Ranch alley, fresh copied checkpoint per process; interleaved explicit1/unset pairs then rollback,NGX and cold-DLAA controls; FG phases1250+,other phases650+ presents | Run dlaa-unset-h3: 1 phases recorded, all stable play; normal exit. AMD [made, requests, missing]: not enabled; last GPU-time samples by phase (ms): dlaa: 10.4. Capped functional/parity run, not an uncapped performance or equivalence claim; no device loss. |
| 2026-11-14 | rollout3fa9b801, combined prerequisites757ba924; isolated renderer pack on unchanged development base514591a3 with common scene matcher, upscaler and FG fixes;19 relevant production functions/helpers match current source; base pipeline-creation wrappers retained | Windows kernel26300 /26H2, windowed2560x1421 client | RTX4070 12GB, NVIDIA621.01 | Core i7-13700F,32GB | 2560x1440 output; RX_VK_PSYCHOV35_DLSS=1; DLAA from saved Native choice, FG Off; Native HDR PsychoV35; paper white200/peak1000,sharpness0,cap61 | Daylight Lon Lon Ranch alley, fresh copied checkpoint per process; interleaved explicit1/unset pairs then rollback,NGX and cold-DLAA controls; FG phases1250+,other phases650+ presents | Run dlaa-explicit1-g3: 1 phases recorded, all stable play; normal exit. AMD [made, requests, missing]: not enabled; last GPU-time samples by phase (ms): dlaa: 10.6. Capped functional/parity run, not an uncapped performance or equivalence claim; no device loss. |
| 2026-11-14 | rollout3fa9b801, combined prerequisites757ba924; isolated renderer pack on unchanged development base514591a3 with common scene matcher, upscaler and FG fixes;19 relevant production functions/helpers match current source; base pipeline-creation wrappers retained | Windows kernel26300 /26H2, windowed2560x1421 client | RTX4070 12GB, NVIDIA621.01 | Core i7-13700F,32GB | 2560x1440 output; RX_VK_PSYCHOV35_DLSS=0; FSR3 at67%, PsychoV35 with AMD FG2x then Vanilla with FG Off; paper white200/peak1000,sharpness0,cap61 | Daylight Lon Lon Ranch alley, fresh copied checkpoint per process; interleaved explicit1/unset pairs then rollback,NGX and cold-DLAA controls; FG phases1250+,other phases650+ presents | Run disable0-e3: 2 phases recorded, all stable play; normal exit. AMD [made, requests, missing]: fsr67 [0, 600, 600]; last GPU-time samples by phase (ms): fsr67: 9.3, vanilla67: 8.5. Capped functional/parity run, not an uncapped performance or equivalence claim; no device loss. |
| 2026-11-14 | host and game folder built from main 9c9d69ed (v2.1.0-rc.3-463-g9c9d69ed); longshot_vk.so built alone from main 414f56e2 (nothing under gpu/ differs from 9c9d69ed), once without a play profile and once with the committed one (RX_VK_PGO= the unpacked assets/profiles/longshot_vk.profdata.xz; the build log says built with the profile), same flags otherwise, the Linux builder’s clang (LLVM 23.1.5) |
SteamOS 3.9.5 (kernel 7.2.7-valve2), headless | Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel; the GPU’s clock fixed at 1600 MHz (Steam’s performance menu) | AMD Custom APU 0932 (4 cores, 8 threads), 16 GB shared with the GPU | Minimum preset (RX_VK_LOW_SPEC=1), handheld profile (HOS_HANDHELD=1), the game’s own upscaler, RX_VK_VRAM_MB=2400, RX_VK_TEX_MAX_DIM=1024, RX_VK_LOAD=create, RX_VK_READBACK_LATE=newest,3980, RX_VK_DESC_POOL_DIV=4, HOS_RESERVE_CORE=0; no mesh shaders: task and mesh programs run as compute shaders, the path the profile was not recorded on |
a save loaded by script, standing still: a heavy view (about 2,150 draws a frame) | p0lux run01, 720 presents measured a run, 2 runs each, alternated with two other renderers, the second round in the reverse order; and the profile’s renderer’s warm-up start. Without the profile against with it: 28.62 / 28.46 fps against 30.68 / 32.30 (31.11 at the warm-up); replay 31.12 / 31.17 ms against 26.99 / 26.50 ms (26.49); the render thread’s instructions a frame 62.38 / 61.99 M against 56.19 / 56.04 M (56.13), its cycles 79.15 / 78.77 M against 68.99 / 68.76 M (68.23). The shaders prepared for one renderer are read by the other (47,797 program keys kept). No errors in 5 starts |
| 2026-11-14 | test/play af15c4a5 (main f3fe1763 with opt-in branches that change no shaders), the same build in both arms; its longshot_pipeline_states_shipped.bin merged on main 7ff5256e against the contributed tour list (contrib/gumminaomi-pipeline-tour f5e7c33f) |
Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | the game’s default options (Auto preset: DLSS, 2048x1152, 60 fps), RX_VK_LOAD=create; no notes and no pipeline cache before each round; the driver’s shader cache warm |
the main outdoor area at night from a save; two interleaved rounds (main, list, main, list): longshot --prepare-shaders --compile-shaders most, then a first and a later start, each 30 s into play |
Noted pipelines 24,464 / 24,470 against 34,352 / 34,912; the build step 17.8 / 22.5 s against 20.0 / 20.5 s; first start’s precompile 2.7 / 2.4 s against 3.4 / 3.6 s; later start’s loads at create 2.8 / 4.2 s against 4.5 / 4.3 s, launch to play 26.1 / 27.5 s against 27.8 / 27.5 s; compute programs kept ready 234 against 2,199; private memory in play 22,877 / 22,927 MB against 23,638 / 24,770 MB. Every pipeline built, no errors |
| 2026-11-14 | this change rebased: 70297a11 on main 26a2eab0 (pack movie-pace-v2, longshot v3.0.0-rc.3-69-g70297a11) |
SteamOS holo, kernel 6.18.50-valve2, gamescope session | AMD Custom GPU 0405 (RADV VANGOGH), Mesa 26.1.2.221562, driver 0x6801063 | Steam Deck OLED (Angela Morley), 4 cores / 8 threads, 16 GB | Game frame rate cap row at 60 (frame_rate_cap=0), vsync off, HOS_VIDEO_LOG=1, HOS_VIDEO_INPUT_LIMIT=0, no HOS_FPS_MAX, FSR 3.1, this change’s movie pacing on (its default) |
the same Deck, the same title screen’s looped intro movie of 5,947 pictures, one launch, 2 plays | 538 / 436 pictures skipped against main’s median 1,043; vi: a movie is playing: the frame rate goes to 240 for it at each movie, vi: the movie is over: back to the game's frame rate cap 60 between them (8 of them); the picture behind by -0.04 s; run03 |
| 2026-11-14 | this change d3a75bf4 on main cf605c85 (longshot v3.0.0-rc.3-60-gd3a75bf4, renderer longshot_vk.so sha256 590f623abc34e3305bbd18a3bfaeb0503c84dd6e3d017cb636acce94c3f14ebc) |
SteamOS holo, kernel 6.18.50-valve2, gamescope session | AMD Custom GPU 0405 (RADV VANGOGH), Mesa 26.1.2.221562, driver 0x6801063 | Steam Deck OLED (Angela Morley), 4 cores / 8 threads, 16 GB | Game frame rate cap row at 60 (frame_rate_cap=0), vsync off, HOS_VIDEO_LOG=1, HOS_VIDEO_INPUT_LIMIT=0, no HOS_FPS_MAX, FSR 3.1, this change’s movie pacing on (its default) |
the same Deck, the same title screen’s looped intro movie of 5,947 pictures, warm caches, 2 plays in one launch; an earlier launch paid the pack’s shader re-prepare and measured 626 / 464 | 551 / 460 pictures skipped (median 506, against main’s median 1,043); vi: a movie is playing: the frame rate goes to 240 for it at each movie and vi: the movie is over: back to the game's frame rate cap 60 between them; the picture behind by -0.02 s; run02 |
| 2026-11-14 | main cf605c85 (longshot v3.0.0-rc.3-59-gcf605c85, renderer longshot_vk.so sha256 9d98436ac449b067ad1226a472c5a9dec2c71b538bd509bed58f87aea85ff9fd) |
SteamOS holo, kernel 6.18.50-valve2, gamescope session | AMD Custom GPU 0405 (RADV VANGOGH), Mesa 26.1.2.221562, driver 0x6801063 | Steam Deck OLED (Angela Morley), 4 cores / 8 threads, 16 GB | Game frame rate cap row at 60 (frame_rate_cap=0), vsync off, HOS_VIDEO_LOG=1, HOS_VIDEO_INPUT_LIMIT=0, no HOS_FPS_MAX, FSR 3.1 |
the title screen’s looped intro movie of 5,947 pictures, warm caches, 4 plays in one launch | 1,182 / 1,074 / 1,012 / 1,028 pictures skipped to keep up with its sound (median 1,043); the picture on the movie’s clock (behind by 0.00 s); no pacing log line; run01 |
| 2026-11-14 | brett-d/legendary-broccoli 52a61f40 (Longshot main 25ee6a83 + this change) built from source on Linux (clang), against main 25ee6a83 (longshot_vk.so swapped in) |
Pop!_OS 24.04 (COSMIC, Wayland), headless | NVIDIA GeForce RTX 3070 Ti 8 GB (also drives the desktop), driver 595.104.02 | Intel i7-12700K, 16 GB RAM (15.7 GB usable), 19 GB swap | Quality preset Auto (this PC), default settings (RX_VK_LOAD=create, 2 rounds of 3 areas) and RX_VK_LOAD=start (first start + 3 areas) |
Saves in the starting forest, the main outdoor area and the village (child), scripted walk with camera turns, 150 s a run (test request T-B56B, 24 runs) | Caches kept across update (47866 program keys kept, 0 re-prepared). Default RX_VK_LOAD=create (median of 6 runs across the 3 areas): peak RSS 12,248 MB (main) vs 11,439 MB (PR, -809 MB), MemAvailable minimum 258 MB (main) vs 643 MB (PR, +385 MB), 1 s samples < 500 MB 141 (main) vs 41 (PR), fps 52.7 (main) vs 54.4 (PR). With RX_VK_LOAD=start: PR lets go of 35,338-35,565 shader modules after the precompile and cuts process swap from 7,945-8,081 MB (main) to 4,940-5,256 MB (PR, ~2.9 GB lower). 0 OOM kills, 0 device losses, 0 remake failures across 24 runs |
| 2026-11-14 | this change (Auto the default) on main cf605c85 with four changes not merged yet: the Battlemage float-add fix (#444), the uniform snapshot on Arc discrete cards (#446), the split-before-draw rule (#443) and XeSS (#459); the renderer built from source into a nightly-80-7315957c game folder never prepared (RX_VK_PRECOMPILE=0) |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | RX_VK_NO_INTEL_SERIAL=0 (One at a time; the log: entered by one thread); frame generation off |
one launch through the Fire Temple and the Water Temple, each by the debug teleport for the first time (a third teleport, to the Spirit Temple, was not taken); new programs counted from the log’s vk: in play: lines, the frame rate from its 5 s vi: windows and from the control bridge’s status every 3 to 5 s |
372 new programs (65, 298, 9 more), the picture under 5 fps for 181 s, 25 fps or more after 73 and 154 s; no device loss, no fault line (run U2) |
| 2026-11-14 | this change (Auto the default) on main cf605c85 with four changes not merged yet: the Battlemage float-add fix (#444), the uniform snapshot on Arc discrete cards (#446), the split-before-draw rule (#443) and XeSS (#459); the renderer built from source into a nightly-80-7315957c game folder never prepared (RX_VK_PRECOMPILE=0) |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | Intel shader compiles: Auto, the default (the log: entered side by side); frame generation off | one launch through Death Mountain Crater, Inside Jabu-Jabu’s Belly and the Ice Cavern, each by the debug teleport for the first time; new programs counted from the log’s vk: in play: lines, the frame rate from its 5 s vi: windows and from the control bridge’s status every 3 to 5 s |
373 new programs (73, 175, 125), no 5 s window under 5 fps, 25 fps or more after 21, 27 and 39 s; no device loss, no fault line (run U1) |
| 2026-11-14 | main cf605c85 with four changes not merged yet: the Battlemage float-add fix (#444), the uniform snapshot on Arc discrete cards (#446), the split-before-draw rule (#443) and XeSS (#459); the renderer built from source into a nightly-80-7315957c game folder never prepared (RX_VK_PRECOMPILE=0) |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | Side by side (RX_VK_NO_INTEL_SERIAL=1); frame generation off |
one launch through the water region, the water region, Death Mountain Trail, the Lost Woods, Gerudo Valley, Inside the first dungeon, Dodongo’s Cavern and the Desert Colossus, each by the debug teleport for the first time; new programs counted from the log’s vk: in play: lines, the frame rate from its 5 s vi: windows and from the control bridge’s status every 3 to 5 s |
850 new programs in all (182, 199, 82, 12, 132, 77, 99, 67), 25 fps or more after 80, 45, 24, 20, 26, 26, 27 and 24 s, never polled under 5 fps; no device loss, no fault line (run T1) |
| 2026-11-14 | main cf605c85 with four changes not merged yet: the Battlemage float-add fix (#444), the uniform snapshot on Arc discrete cards (#446), the split-before-draw rule (#443) and XeSS (#459); the renderer built from source into a nightly-80-7315957c game folder never prepared (RX_VK_PRECOMPILE=0) |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | Side by side (RX_VK_NO_INTEL_SERIAL=1), RX_VK_TRANSLATE_WAIT_MS=8; frame generation off |
the mountain settlement, by the debug teleport for the first time; new programs counted from the log’s vk: in play: lines, the frame rate from its 5 s vi: windows and from the control bridge’s status every 3 to 5 s |
106 new programs, no 5 s window under 5 fps, 25 fps or more after 30 s; no device loss, no fault line (run F3) |
| 2026-11-14 | main cf605c85 with four changes not merged yet: the Battlemage float-add fix (#444), the uniform snapshot on Arc discrete cards (#446), the split-before-draw rule (#443) and XeSS (#459); the renderer built from source into a nightly-80-7315957c game folder never prepared (RX_VK_PRECOMPILE=0) |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | Side by side (RX_VK_NO_INTEL_SERIAL=1); frame generation off |
the village, by the debug teleport for the first time; new programs counted from the log’s vk: in play: lines, the frame rate from its 5 s vi: windows and from the control bridge’s status every 3 to 5 s |
155 new programs, no 5 s window under 5 fps (the lowest 9.0), 25 fps or more after 52 s; no device loss, no fault line (run F2) |
| 2026-11-14 | main cf605c85 with four changes not merged yet: the Battlemage float-add fix (#444), the uniform snapshot on Arc discrete cards (#446), the split-before-draw rule (#443) and XeSS (#459); the renderer built from source into a nightly-80-7315957c game folder never prepared (RX_VK_PRECOMPILE=0) |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | One at a time, RX_VK_TRANSLATE_WAIT_MS=8; frame generation off |
the Graveyard, by the debug teleport for the first time; new programs counted from the log’s vk: in play: lines, the frame rate from its 5 s vi: windows and from the control bridge’s status every 3 to 5 s |
134 new programs, the picture under 5 fps for 73 s, 25 fps or more after 134 s (run F1) |
| 2026-11-14 | main cf605c85 with four changes not merged yet: the Battlemage float-add fix (#444), the uniform snapshot on Arc discrete cards (#446), the split-before-draw rule (#443) and XeSS (#459); the renderer built from source into a nightly-80-7315957c game folder never prepared (RX_VK_PRECOMPILE=0) |
Windows 11 Pro 10.0.29704 | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K (20 threads, 6 background workers), 64 GB | Intel shader compiles: One at a time (main’s default); frame generation off | the main outdoor area then the Market, each by the debug teleport for the first time on this PC; new programs counted from the log’s vk: in play: lines, the frame rate from its 5 s vi: windows and from the control bridge’s status every 3 to 5 s |
the main outdoor area: 398 new programs, the picture under 5 fps for 248 s; the Market: 322 more, 409 s; no device loss, no fault line (run A0) |
| 2026-11-14 | v3.0.0-rc.3-85-g30d3b3c7 (feat/hud-layout on main 18c79f4f, llvm-mingw, NVIDIA DLSS SDK) against v3.0.0-rc.3-84-g8e05abb0 (its parent: the icon at the screen’s corner), one run each per shape and mode, on the same game files | Windows 11 Pro (26300), headless (offscreen, scripted input through the control bridge) | RTX 5080 16 GB, 621.24 | Ryzen 9 9900X3D, 64 GB | HOS_OUTPUT_RES 3440x1440 and 5120x1440, Ultrawide menus blurred (1) and black (2) sides, pictures through the scale pass (RX_VK_SHOTS, RX_VK_SHOT_FX=1, every 6th present), no sound, HOS_FPS_MAX=60, HOS_NO_PACING=1 |
Walking out of the starting forest into the loading fade | Before, 3440x1440 black sides: the corner icon gone in every picture while the sides were on (~500 presents). After: the icon in the middle 16:9’s corner in every picture with the sides on: black 92/92 (21:9), 97/97 (32:9); blurred 99/99 (21:9), 97/97 (32:9) |
| 2026-11-14 | v3.0.0-rc.3-52-g11632fec (feat/hud-layout on main 5ad0f503, llvm-mingw, NVIDIA DLSS SDK) against v2.1.0-rc.3-330-g82de567a (main), one run each per shape, on the same game files | Windows 11 Pro (26300), headless (offscreen, scripted input through the control bridge) | RTX 5080 16 GB, 621.24 | Ryzen 9 9900X3D, 64 GB | HOS_OUTPUT_RES 1280x1024 (5:4), 1366x1000, 1440x1080 (4:3), 1920x1280 (3:2), 1920x1200 (16:10), 1920x1080, 2560x1080 and 3440x1440 (21:9), 3840x1600 (12:5), 5120x1440 (32:9), 5760x1080 (48:9), and 1080x1920 asked for (refused: the game’s own 2048x1152), Ultrawide menus Centred, no sound, HOS_NO_PACING=1 |
the starting forest walked out to the main outdoor area, a shot every 2 presents through the loading fade | The loading icon whole in the screen’s bottom-right corner at every shape from 5:4 to 12:5, at 12:5’s edge at 32:9 and in the middle screen’s corner at 48:9, the fade black to the screen’s edges at all of them. The head before (7cedec2d) cut the icon off at 4:3; main squeezes it below 16:9 and leaves the sides lit above it (run04) |
| 2026-11-14 | longshot v3.0.0-rc.3-59-gcf605c85 (source 8672daee): main cf605c85 as a dev runtime with no pack (the A/B’s main side) | Windows 11 Pro, fullscreen, 5120x2160 output | RTX 5090 32 GB, NVIDIA 621.24 | Ryzen 7 9800X3D, 64 GB | DLAA, DLSS Frame Generation 2x (NGX), V-Sync Auto (FIFO), Reflex On + Boost, frame pacing on | The four Gossip Stones in front of the Temple of Time, the camera moved back and forth (the PR’s repro), about a minute, judged by eye; run 4 of the A/B’s 4, alternated (this PR, main, this PR, main) | run04 (main): the stones break apart in made frames; actor motion 0.056 ms CPU a frame, 1.08 us a skinned draw (median of 2 steady blocks, 51.3 skinned draws a frame) |
| 2026-11-14 | longshot v3.0.0-rc.3-59-gcf605c85 (source 8672daee) with Longshot PR 59’s four commits at cc7cb36a rebased onto it (no conflict), its renderer built as a dev pack on the same base | Windows 11 Pro, fullscreen, 5120x2160 output | RTX 5090 32 GB, NVIDIA 621.24 | Ryzen 7 9800X3D, 64 GB | DLAA, DLSS Frame Generation 2x (NGX), V-Sync Auto (FIFO), Reflex On + Boost, frame pacing on | The four Gossip Stones in front of the Temple of Time, the camera moved back and forth (the PR’s repro), about a minute, judged by eye; run 3 of the A/B’s 4, alternated (this PR, main, this PR, main) | run03 (this PR): the stones stay whole in made frames; poses all read at the draw, 0 copied later; actor motion 0.151 ms CPU a frame, 2.80 us a skinned draw (median of 3 steady blocks, 53.8 skinned draws a frame) |
| 2026-11-14 | longshot v3.0.0-rc.3-59-gcf605c85 (source 8672daee): main cf605c85 as a dev runtime with no pack (the A/B’s main side) | Windows 11 Pro, fullscreen, 5120x2160 output | RTX 5090 32 GB, NVIDIA 621.24 | Ryzen 7 9800X3D, 64 GB | DLAA, DLSS Frame Generation 2x (NGX), V-Sync Auto (FIFO), Reflex On + Boost, frame pacing on | The four Gossip Stones in front of the Temple of Time, the camera moved back and forth (the PR’s repro), about a minute, judged by eye; run 2 of the A/B’s 4, alternated (this PR, main, this PR, main) | run02 (main): the stones break apart in made frames; actor motion 0.055 ms CPU a frame, 1.09 us a skinned draw (median of 7 steady blocks, 50.5 skinned draws a frame) |
| 2026-11-14 | longshot v3.0.0-rc.3-59-gcf605c85 (source 8672daee) with Longshot PR 59’s four commits at cc7cb36a rebased onto it (no conflict), its renderer built as a dev pack | Windows 11 Pro, fullscreen, 5120x2160 output | RTX 5090 32 GB, NVIDIA 621.24 | Ryzen 7 9800X3D, 64 GB | DLAA, DLSS Frame Generation 2x (NGX), V-Sync Auto (FIFO), Reflex On + Boost, frame pacing on | The four Gossip Stones in front of the Temple of Time, the camera moved back and forth (the PR’s repro); one run judged by eye, numbers from 8 status blocks of 600 frames in play | run01: the stones stay whole in made frames; DLSS-G made 600 of 600 presents in 7 of 8 blocks (541 in the last); poses all read at the draw, 0 copied later; actor motion 0.139 ms CPU a frame (median of the 6 steady blocks, 0.133-0.178 ms) |
| 2026-11-14 UTC | v3.0.0-rc.3, guest memory reporting backported from a4c77174, host db2f0d91f2b7, renderer bb5fd0cf939d | SteamOS 3.9.5, direct headless | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, copied warm caches, HOS_MEMINFO_OLD=0, RX_TEST_PARTIAL_PROBE=1 | Child starting-area save, 20 s settled, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warmup and 45 s stationary. | partial. Field reached after 8.0 s. 20.08 FPS, p95 interval 53.33 ms. Cumulative probe from process start: 0 of 761548800 queried bytes outside the conservatively estimated import coverage, 2770 uncertain queries counted as fully covered. Span totals count repeated addresses and do not measure reclaimed RAM. |
| 2026-11-14 UTC | v3.0.0-rc.3, guest memory reporting backported from a4c77174, host db2f0d91f2b7, renderer 7f8360cf60c2 | SteamOS 3.9.5, direct headless | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, copied warm caches, HOS_MEMINFO_OLD=0, RX_TEST_PARTIAL_PROBE=1 | Child starting-area save, 20 s settled, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warmup and 45 s stationary. | partial-v2. Field reached after 8.0 s. 20.40 FPS, p95 interval 52.85 ms. Cumulative probe from process start: 19083264 of 962428928 queried bytes outside the conservatively estimated import coverage, 0 uncertain queries counted as fully covered. Span totals count repeated addresses and do not measure reclaimed RAM. |
| 2026-11-14 UTC | v3.0.0-rc.3, guest memory reporting backported from a4c77174, host db2f0d91f2b7, renderer ce5646666a7e | SteamOS 3.9.5, direct headless | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, copied warm caches, HOS_MEMINFO_OLD=0, RX_TEST_IMPORT_EXCLUSIVE_END=0 | Child starting-area save, 20 s settled, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warmup and 45 s stationary. | import-old. Field reached after 8.0 s. 20.09 FPS, p95 interval 53.44 ms. The probe logged exactly two adjacent-only frees during transition, 8 and 64 KiB, with real overlaps preserved. |
| 2026-11-14 UTC | v3.0.0-rc.3, guest memory reporting backported from a4c77174, host db2f0d91f2b7, renderer ce5646666a7e | SteamOS 3.9.5, direct headless | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, copied warm caches, HOS_MEMINFO_OLD=0, RX_TEST_IMPORT_EXCLUSIVE_END=1 | Child starting-area save, 20 s settled, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warmup and 45 s stationary. | import-fixed. Field reached after 9.0 s. 20.33 FPS, p95 interval 52.36 ms. The probe logged exactly two adjacent-only frees during transition, 8 and 64 KiB, with real overlaps preserved. |
| 2026-11-14 UTC | v3.0.0-rc.3, guest memory reporting backported from a4c77174, host db2f0d91f2b7, renderer 9a43e860629d | SteamOS 3.9.5, direct headless | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, copied warm caches, HOS_MEMINFO_OLD=0, RX_VK_PASS_TIMING=2, RX_VK_PROF_FINE=1 | Child starting-area save, 20 s settled, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warmup and 45 s stationary. | field-gpu. Field reached after 8.0 s. 19.54 FPS, p95 interval 55.25 ms. Timing-enabled 600-present report in Field averaged 27.16 ms GPU time, with 8.04 ms in compute dispatches. Instrumentation affects timing. |
| 2026-11-14 UTC | v3.0.0-rc.3, guest memory reporting backported from a4c77174, host db2f0d91f2b7, renderer 9a43e860629d | SteamOS 3.9.5, direct headless | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, copied warm caches, HOS_MEMINFO_OLD=0, HOS_RESERVE_CORE=0 | Child starting-area save, 20 s settled, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warmup and 45 s stationary. | affinity-off. Field reached after 8.0 s. 20.73 FPS, p95 interval 51.45 ms. Exploratory single run with core reservation disabled. |
| 2026-11-14 UTC | v3.0.0-rc.3 (release source fingerprint 51559a15) with the guest memory reporting portion of a4c77174 backported, test fingerprint 9ded2fb4, stock renderer | SteamOS 3.9.5 build 20261023.100, direct headless run | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared RAM | 1280×720 output, render resolution option 4, frame generation off, warm cache copied per launch, 1536 MiB texture budget, RX_VK_PRECOMPILE=0, HOS_MEMINFO_OLD=0, RX_VK_RECORDER=thread | Fresh copy of child starting-area save, 20 s settled at spawn, teleport to HyruleField_Child. Available RAM sampled every 0.5 s, 768 MiB stop threshold or sustained full pressure below 1 GiB. | E3. Verified Field gameplay after 8.0 s. After 20 s warmup, 45 s standing yielded 20.08 FPS, p95 interval 53.6 ms. Minimum available system memory from accepted teleport to stop or sample end 1598 MiB. Last/highest logged mapped-page peak 2560 MiB. Guest memfd backing at settled starting-area 5405 MiB. After a confirmed deep suspend and Bluetooth wake, GPU policy auto. |
| 2026-11-14 UTC | v3.0.0-rc.3 (release source fingerprint 51559a15) with the guest memory reporting portion of a4c77174 backported, test fingerprint 9ded2fb4, stock renderer | SteamOS 3.9.5 build 20261023.100, direct headless run | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared RAM | 1280×720 output, render resolution option 4, frame generation off, warm cache copied per launch, 1536 MiB texture budget, RX_VK_PRECOMPILE=0, HOS_MEMINFO_OLD=0, RX_VK_RECORDER=thread | Fresh copy of child starting-area save, 20 s settled at spawn, teleport to HyruleField_Child. Available RAM sampled every 0.5 s, 768 MiB stop threshold or sustained full pressure below 1 GiB. | E1. Verified Field gameplay after 8.0 s. After 20 s warmup, 45 s standing yielded 20.31 FPS, p95 interval 53.2 ms. Minimum available system memory from accepted teleport to stop or sample end 1680 MiB. Last/highest logged mapped-page peak 2560 MiB. Guest memfd backing at settled starting-area 5404 MiB. After a confirmed deep suspend and Bluetooth wake, GPU policy auto. Post-sample screenshot failed in the harness. The game exited normally. |
| 2026-11-14 UTC | v3.0.0-rc.3 (release source fingerprint 51559a15) with the guest memory reporting portion of a4c77174 backported, test fingerprint 9ded2fb4, stock renderer | SteamOS 3.9.5 build 20261023.100, direct headless run | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared RAM | 1280×720 output, render resolution option 4, frame generation off, warm cache copied per launch, 1536 MiB texture budget, RX_VK_PRECOMPILE=0, HOS_MEMINFO_OLD=1 | Fresh copy of child starting-area save, 20 s settled at spawn, teleport to HyruleField_Child. Available RAM sampled every 0.5 s, 768 MiB stop threshold or sustained full pressure below 1 GiB. | D1. Low-memory guard stopped the game 6.1 s after teleport acceptance, before verified Field gameplay. Minimum available system memory from accepted teleport to stop or sample end 676 MiB. Last/highest logged mapped-page peak 3594 MiB. Guest memfd backing at settled starting-area 6998 MiB. After a confirmed deep suspend and Bluetooth wake, GPU policy auto. |
| 2026-11-14 UTC | v3.0.0-rc.3 (release source fingerprint 51559a15) with the guest memory reporting portion of a4c77174 backported, test fingerprint 9ded2fb4, stock renderer | SteamOS 3.9.5 build 20261023.100, direct headless run | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared RAM | 1280×720 output, render resolution option 4, frame generation off, warm cache copied per launch, 1536 MiB texture budget, RX_VK_PRECOMPILE=0, HOS_MEMINFO_OLD=0 | Fresh copy of child starting-area save, 20 s settled at spawn, teleport to HyruleField_Child. Available RAM sampled every 0.5 s, 768 MiB stop threshold or sustained full pressure below 1 GiB. | C3. Verified Field gameplay after 8.0 s. After 20 s warmup, 45 s standing yielded 20.44 FPS, p95 interval 52.4 ms. Minimum available system memory from accepted teleport to stop or sample end 1493 MiB. Last/highest logged mapped-page peak 2560 MiB. Guest memfd backing at settled starting-area 5411 MiB. After a confirmed deep suspend and Bluetooth wake, GPU policy auto. |
| 2026-11-14 UTC | v3.0.0-rc.3 (release source fingerprint 51559a15) with the guest memory reporting portion of a4c77174 backported, test fingerprint 9ded2fb4, stock renderer | SteamOS 3.9.5 build 20261023.100, direct headless run | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared RAM | 1280×720 output, render resolution option 4, frame generation off, warm cache copied per launch, 1536 MiB texture budget, RX_VK_PRECOMPILE=0, HOS_MEMINFO_OLD=0 | Fresh copy of child starting-area save, 20 s settled at spawn, teleport to HyruleField_Child. Available RAM sampled every 0.5 s, 768 MiB stop threshold or sustained full pressure below 1 GiB. | C2. Verified Field gameplay after 9.0 s. After 20 s warmup, 45 s standing yielded 20.20 FPS, p95 interval 53.4 ms. Minimum available system memory from accepted teleport to stop or sample end 1774 MiB. Last/highest logged mapped-page peak 2560 MiB. Guest memfd backing at settled starting-area 5434 MiB. After a confirmed deep suspend and Bluetooth wake, GPU policy auto. |
| 2026-11-14 UTC | v3.0.0-rc.3 (release source fingerprint 51559a15) with the guest memory reporting portion of a4c77174 backported, test fingerprint 9ded2fb4, stock renderer | SteamOS 3.9.5 build 20261023.100, direct headless run | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared RAM | 1280×720 output, render resolution option 4, frame generation off, warm cache copied per launch, 1536 MiB texture budget, RX_VK_PRECOMPILE=0, HOS_MEMINFO_OLD=0 | Fresh copy of child starting-area save, 20 s settled at spawn, teleport to HyruleField_Child. Available RAM sampled every 0.5 s, 768 MiB stop threshold or sustained full pressure below 1 GiB. | C1. Verified Field gameplay after 9.0 s. After 20 s warmup, 45 s standing yielded 20.47 FPS, p95 interval 52.0 ms. Minimum available system memory from accepted teleport to stop or sample end 1743 MiB. Last/highest logged mapped-page peak 2569 MiB. Guest memfd backing at settled starting-area 5418 MiB. After a confirmed deep suspend and Bluetooth wake, GPU policy auto. |
| 2026-11-14 UTC | v3.0.0-rc.3 (release source fingerprint 51559a15) with the guest memory reporting portion of a4c77174 backported, test fingerprint 9ded2fb4, stock renderer | SteamOS 3.9.5 build 20261023.100, direct headless run | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared RAM | 1280×720 output, render resolution option 4, frame generation off, warm cache copied per launch, 1536 MiB texture budget, RX_VK_PRECOMPILE=0, HOS_MEMINFO_OLD=0 | Fresh copy of child starting-area save, 20 s settled at spawn, teleport to HyruleField_Child. Available RAM sampled every 0.5 s, 768 MiB stop threshold or sustained full pressure below 1 GiB. | B2. Verified Field gameplay after 14.0 s. After 20 s warmup, 45 s standing yielded 6.03 FPS, p95 interval 179.5 ms. Minimum available system memory from accepted teleport to stop or sample end 1141 MiB. Last/highest logged mapped-page peak 2818 MiB. Guest memfd backing at settled starting-area 5414 MiB. |
| 2026-11-14 UTC | v3.0.0-rc.3 (release source fingerprint 51559a15) with the guest memory reporting portion of a4c77174 backported, test fingerprint 9ded2fb4, stock renderer | SteamOS 3.9.5 build 20261023.100, direct headless run | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared RAM | 1280×720 output, render resolution option 4, frame generation off, warm cache copied per launch, 1536 MiB texture budget, RX_VK_PRECOMPILE=0, HOS_MEMINFO_OLD=0 | Fresh copy of child starting-area save, 20 s settled at spawn, teleport to HyruleField_Child. Available RAM sampled every 0.5 s, 768 MiB stop threshold or sustained full pressure below 1 GiB. | B1. Verified Field gameplay after 14.0 s. After 20 s warmup, 45 s standing yielded 6.13 FPS, p95 interval 176.4 ms. Minimum available system memory from accepted teleport to stop or sample end 1280 MiB. Last/highest logged mapped-page peak 2568 MiB. Guest memfd backing at settled starting-area 5472 MiB. |
| 2026-11-14 UTC | v3.0.0-rc.3 (release source fingerprint 51559a15) with the guest memory reporting portion of a4c77174 backported, test fingerprint 9ded2fb4, stock renderer | SteamOS 3.9.5 build 20261023.100, direct headless run | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared RAM | 1280×720 output, render resolution option 4, frame generation off, warm cache copied per launch, 1536 MiB texture budget, RX_VK_PRECOMPILE=0, HOS_MEMINFO_OLD=1 | Fresh copy of child starting-area save, 20 s settled at spawn, teleport to HyruleField_Child. Available RAM sampled every 0.5 s, 768 MiB stop threshold or sustained full pressure below 1 GiB. | A2. Low-memory guard stopped the game 6.5 s after teleport acceptance, before verified Field gameplay. Minimum available system memory from accepted teleport to stop or sample end 759 MiB. Last/highest logged mapped-page peak 3584 MiB. Guest memfd backing at settled starting-area 6980 MiB. |
| 2026-11-14 UTC | v3.0.0-rc.3 (release source fingerprint 51559a15) with the guest memory reporting portion of a4c77174 backported, test fingerprint 9ded2fb4, stock renderer | SteamOS 3.9.5 build 20261023.100, direct headless run | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared RAM | 1280×720 output, render resolution option 4, frame generation off, warm cache copied per launch, 1536 MiB texture budget, RX_VK_PRECOMPILE=0, HOS_MEMINFO_OLD=1 | Fresh copy of child starting-area save, 20 s settled at spawn, teleport to HyruleField_Child. Available RAM sampled every 0.5 s, 768 MiB stop threshold or sustained full pressure below 1 GiB. | A1. Low-memory guard stopped the game 6.5 s after teleport acceptance, before verified Field gameplay. Minimum available system memory from accepted teleport to stop or sample end 702 MiB. Last/highest logged mapped-page peak 3587 MiB. Guest memfd backing at settled starting-area 6979 MiB. |
| 2026-11-14 UTC | #469 @ 76382425 (renderer v3.0.0-rc.3-78-g76382425) on main 9c9d69ed: dev_packs renderer pack, host identical to main’s | SteamOS 3.9.5, direct headless (HOS_HEADLESS=1) | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | release optimisation (O2, profile), x86-64-v3; 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, 256 MiB mirror budget; caches from one shared warm-up pass of main copied fresh into every run | Child starting-area save settled 20 s, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warm-up and 45 s stationary; one session interleaving main, #468, #469 and #472 in that order, four rounds (round 3’s #472 run, cut by a Deck suspend, is discarded; round 4 ran after the resume); present intervals from HOS_PRESENT_TRACE in the 45 s window | pr469-r4, round 4. Field reached 11.7 s after the teleport. 20.38 FPS, present interval median 49.05 ms, p95 52.05 ms; GPU clock median 1040 MHz at 99% busy. Lowest available memory 1908 MiB, guest memory backed 5900 MiB at the end of the sample; no device fault, no run guard, exit 0. |
| 2026-11-14 UTC | #469 @ 76382425 (renderer v3.0.0-rc.3-78-g76382425) on main 9c9d69ed: dev_packs renderer pack, host identical to main’s | SteamOS 3.9.5, direct headless (HOS_HEADLESS=1) | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | release optimisation (O2, profile), x86-64-v3; 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, 256 MiB mirror budget; caches from one shared warm-up pass of main copied fresh into every run | Child starting-area save settled 20 s, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warm-up and 45 s stationary; one session interleaving main, #468, #469 and #472 in that order, four rounds (round 3’s #472 run, cut by a Deck suspend, is discarded; round 4 ran after the resume); present intervals from HOS_PRESENT_TRACE in the 45 s window | pr469-r3, round 3. Field reached 7.2 s after the teleport. 20.52 FPS, present interval median 48.69 ms, p95 51.99 ms; GPU clock median 1040 MHz at 99% busy. Lowest available memory 1991 MiB, guest memory backed 5845 MiB at the end of the sample; no device fault, no run guard, exit 0. |
| 2026-11-14 UTC | #469 @ 76382425 (renderer v3.0.0-rc.3-78-g76382425) on main 9c9d69ed: dev_packs renderer pack, host identical to main’s | SteamOS 3.9.5, direct headless (HOS_HEADLESS=1) | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | release optimisation (O2, profile), x86-64-v3; 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, 256 MiB mirror budget; caches from one shared warm-up pass of main copied fresh into every run | Child starting-area save settled 20 s, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warm-up and 45 s stationary; one session interleaving main, #468, #469 and #472 in that order, four rounds (round 3’s #472 run, cut by a Deck suspend, is discarded; round 4 ran after the resume); present intervals from HOS_PRESENT_TRACE in the 45 s window | pr469-r2, round 2. Field reached 9.7 s after the teleport. 20.66 FPS, present interval median 48.39 ms, p95 51.53 ms; GPU clock median 1040 MHz at 94% busy. Lowest available memory 2142 MiB, guest memory backed 5875 MiB at the end of the sample; no device fault, no run guard, exit 0. |
| 2026-11-14 UTC | #469 @ 76382425 (renderer v3.0.0-rc.3-78-g76382425) on main 9c9d69ed: dev_packs renderer pack, host identical to main’s | SteamOS 3.9.5, direct headless (HOS_HEADLESS=1) | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | release optimisation (O2, profile), x86-64-v3; 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, 256 MiB mirror budget; caches from one shared warm-up pass of main copied fresh into every run | Child starting-area save settled 20 s, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warm-up and 45 s stationary; one session interleaving main, #468, #469 and #472 in that order, four rounds (round 3’s #472 run, cut by a Deck suspend, is discarded; round 4 ran after the resume); present intervals from HOS_PRESENT_TRACE in the 45 s window | pr469-r1, round 1. Field reached 10.7 s after the teleport. 20.59 FPS, present interval median 48.56 ms, p95 51.42 ms; GPU clock median 1040 MHz at 90% busy. Lowest available memory 1773 MiB, guest memory backed 5832 MiB at the end of the sample; no device fault, no run guard, exit 0. |
| 2026-11-14 UTC | #468 @ 46b467d0 (renderer v3.0.0-rc.3-77-g46b467d0) on main 9c9d69ed: dev_packs renderer pack, host identical to main’s | SteamOS 3.9.5, direct headless (HOS_HEADLESS=1) | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | release optimisation (O2, profile), x86-64-v3; 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, 256 MiB mirror budget; caches from one shared warm-up pass of main copied fresh into every run | Child starting-area save settled 20 s, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warm-up and 45 s stationary; one session interleaving main, #468, #469 and #472 in that order, four rounds (round 3’s #472 run, cut by a Deck suspend, is discarded; round 4 ran after the resume); present intervals from HOS_PRESENT_TRACE in the 45 s window | pr468-r4, round 4. Field reached 8.4 s after the teleport. 20.30 FPS, present interval median 49.31 ms, p95 52.12 ms; GPU clock median 1040 MHz at 99% busy. Lowest available memory 2085 MiB, guest memory backed 5829 MiB at the end of the sample; no device fault, no run guard, exit 0. |
| 2026-11-14 UTC | #468 @ 46b467d0 (renderer v3.0.0-rc.3-77-g46b467d0) on main 9c9d69ed: dev_packs renderer pack, host identical to main’s | SteamOS 3.9.5, direct headless (HOS_HEADLESS=1) | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | release optimisation (O2, profile), x86-64-v3; 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, 256 MiB mirror budget; caches from one shared warm-up pass of main copied fresh into every run | Child starting-area save settled 20 s, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warm-up and 45 s stationary; one session interleaving main, #468, #469 and #472 in that order, four rounds (round 3’s #472 run, cut by a Deck suspend, is discarded; round 4 ran after the resume); present intervals from HOS_PRESENT_TRACE in the 45 s window | pr468-r3, round 3. Field reached 10.6 s after the teleport. 20.62 FPS, present interval median 48.46 ms, p95 51.62 ms; GPU clock median 1040 MHz at 99% busy. Lowest available memory 2123 MiB, guest memory backed 5863 MiB at the end of the sample; no device fault, no run guard, exit 0. |
| 2026-11-14 UTC | #468 @ 46b467d0 (renderer v3.0.0-rc.3-77-g46b467d0) on main 9c9d69ed: dev_packs renderer pack, host identical to main’s | SteamOS 3.9.5, direct headless (HOS_HEADLESS=1) | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | release optimisation (O2, profile), x86-64-v3; 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, 256 MiB mirror budget; caches from one shared warm-up pass of main copied fresh into every run | Child starting-area save settled 20 s, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warm-up and 45 s stationary; one session interleaving main, #468, #469 and #472 in that order, four rounds (round 3’s #472 run, cut by a Deck suspend, is discarded; round 4 ran after the resume); present intervals from HOS_PRESENT_TRACE in the 45 s window | pr468-r2, round 2. Field reached 9.7 s after the teleport. 20.71 FPS, present interval median 48.26 ms, p95 51.29 ms; GPU clock median 1040 MHz at 99% busy. Lowest available memory 2459 MiB, guest memory backed 5866 MiB at the end of the sample; no device fault, no run guard, exit 0. |
| 2026-11-14 UTC | #468 @ 46b467d0 (renderer v3.0.0-rc.3-77-g46b467d0) on main 9c9d69ed: dev_packs renderer pack, host identical to main’s | SteamOS 3.9.5, direct headless (HOS_HEADLESS=1) | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | release optimisation (O2, profile), x86-64-v3; 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, 256 MiB mirror budget; caches from one shared warm-up pass of main copied fresh into every run | Child starting-area save settled 20 s, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warm-up and 45 s stationary; one session interleaving main, #468, #469 and #472 in that order, four rounds (round 3’s #472 run, cut by a Deck suspend, is discarded; round 4 ran after the resume); present intervals from HOS_PRESENT_TRACE in the 45 s window | pr468-r1, round 1. Field reached 10.7 s after the teleport. 20.44 FPS, present interval median 48.74 ms, p95 52.93 ms; GPU clock median 1040 MHz at 99% busy. Lowest available memory 2520 MiB, guest memory backed 5919 MiB at the end of the sample; no device fault, no run guard, exit 0. |
| 2026-11-14 UTC | main @ 9c9d69ed (v3.0.0-rc.3-75-g9c9d69ed): dev_packs base runtime, no pack | SteamOS 3.9.5, direct headless (HOS_HEADLESS=1) | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | release optimisation (O2, profile), x86-64-v3; 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, 256 MiB mirror budget; caches from one shared warm-up pass of main copied fresh into every run | Child starting-area save settled 20 s, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warm-up and 45 s stationary; one session interleaving main, #468, #469 and #472 in that order, four rounds (round 3’s #472 run, cut by a Deck suspend, is discarded; round 4 ran after the resume); present intervals from HOS_PRESENT_TRACE in the 45 s window | main-r4, round 4. Field reached 10.7 s after the teleport. 20.64 FPS, present interval median 48.47 ms, p95 51.52 ms; GPU clock median 1040 MHz at 99% busy. Lowest available memory 2168 MiB, guest memory backed 5866 MiB at the end of the sample; no device fault, no run guard, exit 0. |
| 2026-11-14 UTC | main @ 9c9d69ed (v3.0.0-rc.3-75-g9c9d69ed): dev_packs base runtime, no pack | SteamOS 3.9.5, direct headless (HOS_HEADLESS=1) | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | release optimisation (O2, profile), x86-64-v3; 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, 256 MiB mirror budget; caches from one shared warm-up pass of main copied fresh into every run | Child starting-area save settled 20 s, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warm-up and 45 s stationary; one session interleaving main, #468, #469 and #472 in that order, four rounds (round 3’s #472 run, cut by a Deck suspend, is discarded; round 4 ran after the resume); present intervals from HOS_PRESENT_TRACE in the 45 s window | main-r3, round 3. Field reached 10.7 s after the teleport. 20.69 FPS, present interval median 48.22 ms, p95 51.88 ms; GPU clock median 1040 MHz at 99% busy. Lowest available memory 2259 MiB, guest memory backed 5868 MiB at the end of the sample; no device fault, no run guard, exit 0. |
| 2026-11-14 UTC | main @ 9c9d69ed (v3.0.0-rc.3-75-g9c9d69ed): dev_packs base runtime, no pack | SteamOS 3.9.5, direct headless (HOS_HEADLESS=1) | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | release optimisation (O2, profile), x86-64-v3; 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, 256 MiB mirror budget; caches from one shared warm-up pass of main copied fresh into every run | Child starting-area save settled 20 s, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warm-up and 45 s stationary; one session interleaving main, #468, #469 and #472 in that order, four rounds (round 3’s #472 run, cut by a Deck suspend, is discarded; round 4 ran after the resume); present intervals from HOS_PRESENT_TRACE in the 45 s window | main-r2, round 2. Field reached 10.7 s after the teleport. 20.54 FPS, present interval median 48.64 ms, p95 51.86 ms; GPU clock median 1040 MHz at 96% busy. Lowest available memory 2202 MiB, guest memory backed 5861 MiB at the end of the sample; no device fault, no run guard, exit 0. |
| 2026-11-14 UTC | main @ 9c9d69ed (v3.0.0-rc.3-75-g9c9d69ed): dev_packs base runtime, no pack | SteamOS 3.9.5, direct headless (HOS_HEADLESS=1) | Steam Deck OLED, RADV VANGOGH, Mesa 26.2.0-devel 5c65d4fc31 | AMD Custom APU 0932, 16 GB shared | release optimisation (O2, profile), x86-64-v3; 1280×720 output, render resolution option 4, frame generation off, precompile off, 1536 MiB texture budget, 256 MiB mirror budget; caches from one shared warm-up pass of main copied fresh into every run | Child starting-area save settled 20 s, teleport to HyruleField_Child at the Lost Woods bridge, 20 s warm-up and 45 s stationary; one session interleaving main, #468, #469 and #472 in that order, four rounds (round 3’s #472 run, cut by a Deck suspend, is discarded; round 4 ran after the resume); present intervals from HOS_PRESENT_TRACE in the 45 s window | main-r1, round 1. Field reached 10.7 s after the teleport. 20.56 FPS, present interval median 48.71 ms, p95 51.71 ms; GPU clock median 1040 MHz at 99% busy. Lowest available memory 1602 MiB, guest memory backed 5909 MiB at the end of the sample; no device fault, no run guard, exit 0. |
| 2026-11-14 | host and game folder built from main 000f19fb; longshot_vk.so built alone from main 514591a3 and from the descriptor pool count’s commit on it, d4b89309, same flags |
SteamOS 3.9.5 (kernel 7.2.7-valve2), headless | Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel; the GPU’s clock fixed at 1600 MHz (Steam’s performance menu) | AMD Custom APU 0932 (4 cores, 8 threads), 16 GB shared with the GPU | Minimum preset (RX_VK_LOW_SPEC=1), handheld profile (HOS_HANDHELD=1), the game’s own upscaler, RX_VK_VRAM_MB=2400, RX_VK_TEX_MAX_DIM=1024, RX_VK_DESC_SLOTGEN=1, RX_VK_LOAD=create, HOS_RESERVE_CORE=0; RX_VK_DESC_POOL_DIV unset, then 2, 4, 8 and 16 |
a save loaded by script, standing still: a heavy view (about 2,100 draws a frame), the starting forest (about 2,550), the main outdoor area (about 1,600) | p0lux run02, 720 presents measured a run. Main then the commit, alternated, 3 runs each on the heavy view and 2 in the starting forest: heavy view 23.99 / 23.99 / 23.89 fps against 24.22 / 24.02 / 24.05, the render thread’s instructions a frame 62.42 / 62.68 / 62.85 M against 62.56 / 63.00 / 62.54 M; the starting forest 22.39 / 22.40 fps against 22.40 / 22.75. Descriptor pools made when a recording’s own ran out, the commit’s renderer, at the full size, a half, a quarter, an eighth and a sixteenth: heavy view 0 (3 runs), 0, 0, 1, 7; the starting forest 0 (2 runs), 0, 1 and 2, 716 and 681, 1,044; the main outdoor area 0, 0, 0, 0, 0. No errors in 27 starts |
| 2026-11-14 | host and game folder built from main 000f19fb; longshot_vk.so built alone from 3a3a6475 (main 514591a3 with the descriptor pool count and other renderer-only changes), one build |
SteamOS 3.9.5 (kernel 7.2.7-valve2), headless | Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel; the GPU’s clock fixed at 1600 MHz (Steam’s performance menu) | AMD Custom APU 0932 (4 cores, 8 threads), 16 GB shared with the GPU | Minimum preset (RX_VK_LOW_SPEC=1), handheld profile (HOS_HANDHELD=1), the game’s own upscaler, RX_VK_VRAM_MB=2400, RX_VK_TEX_MAX_DIM=1024, RX_VK_DESC_SLOTGEN=1, RX_VK_LOAD=create, HOS_RESERVE_CORE=0, RX_VK_READBACK_LATE=newest,3980; RX_VK_DESC_POOL_DIV unset, then 16 |
a save loaded by script, standing still: a heavy view (about 2,150 draws a frame), the starting forest (about 2,600), the main outdoor area (about 1,600) | p0lux run01, one run a scene and a setting, 720 presents measured; the last “recording streams” line of each log. Descriptor pools made when a recording’s own ran out, the switch unset then 16: heavy view 0 then 3, the starting forest 0 then 1,691, the main outdoor area 0 then 0. fps: 29.68 then 29.28, 25.67 then 25.78, 25.14 then 25.09. No errors in 9 starts |
| 2026-11-14 | longshot main 839e85ff (v3.0.0-rc.2-1-g839e85ff) with #415’s fix applied (Codex’s four-file backport, 57f06de8; its renderer only), built from source (no DLSS): the PR arm of #415’s A/B | Windows 11, 1280x720 window (16:9) | RX 9070 XT, AMD 26.10.2 | Ryzen 7 5700X3D, 64 GB | defaults (RX_VK_PCACHE_LOAD unset), the fixed Custom test profile |
the main outdoor area, the standard test: the save loaded, the camera turning at full speed, frame rate unlocked (HOS_FPS_MAX=240), 20 s window; 3 alternating pairs, no warm-ups, all 3 valid; both runtimes started from the same pipeline-cache seed (main cache + 3 play files) |
3 runs: fps 63.19, 63.44, 64.79; CPU 65.47, 63.47, 65.44 ms a frame; worst frame 64, 106, 63 ms |
| 2026-11-14 | longshot main 839e85ff (v3.0.0-rc.2-1-g839e85ff), built from source (no DLSS): the control arm of #415’s A/B | Windows 11, 1280x720 window (16:9) | RX 9070 XT, AMD 26.10.2 | Ryzen 7 5700X3D, 64 GB | defaults (RX_VK_PCACHE_LOAD unset), the fixed Custom test profile |
the main outdoor area, the standard test: the save loaded, the camera turning at full speed, frame rate unlocked (HOS_FPS_MAX=240), 20 s window; 3 alternating pairs, no warm-ups, all 3 valid; both runtimes started from the same pipeline-cache seed (main cache + 3 play files) |
3 runs: fps 61.74, 62.54, 63.09; CPU 65.13, 65.70, 65.21 ms a frame; worst frame 580, 398, 183 ms |
| 2026-11-14 | 80e0e5ee: main cf605c85 with this change, the renderer built from source into a nightly-80-7315957c game folder | Windows 11 Pro 10.0.29704, 2048x1152 output | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the default with the change (float adds the GPU’s own atomic); no other switch | from the start through the control bridge into the save’s the starting forest (child), then standing on the player character’s balcony | no device loss in 3,321 frames, 41-45 fps in the forest (5 s windows); run fadd2, 03:29 |
| 2026-11-14 | main cf605c85, the renderer built from source into a nightly-80-7315957c game folder | Windows 11 Pro 10.0.29704, 2048x1152 output | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the default on main (float adds a compare-exchange loop on this card); no other switch | from the start through the control bridge into the save’s the starting forest (child), then standing on the player character’s balcony | the device was lost at frame 936, as the forest came up, after 898 frames presented (VK_ERROR_DEVICE_LOST); run main2, 03:28 |
| 2026-11-14 | 80e0e5ee: main cf605c85 with this change, the renderer built from source into a nightly-80-7315957c game folder | Windows 11 Pro 10.0.29704, 2048x1152 output | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the default with the change (float adds the GPU’s own atomic); no other switch | from the start through the control bridge into the save’s the starting forest (child), then standing on the player character’s balcony | no device loss in 3,379 frames, 38-45 fps in the forest (5 s windows); run fadd1, 03:27 |
| 2026-11-14 | main cf605c85, the renderer built from source into a nightly-80-7315957c game folder | Windows 11 Pro 10.0.29704, 2048x1152 output | Intel Arc B580 12 GB (PCI device 0xE20B), driver 32.0.101.9154 | Core i5-13600K, 64 GB | the default on main (float adds a compare-exchange loop on this card); no other switch | from the start through the control bridge into the save’s the starting forest (child), then standing on the player character’s balcony | the device was lost at frame 942, as the forest came up, after 898 frames presented (VK_ERROR_DEVICE_LOST, the checkpoints at the dispatch of 499dbfd0721d2858); run main1, 03:26 |
| 2026-11-14 | 47ebdc75 (the uniform snapshot on every Intel GPU, on main cf605c85) with the Battlemage float-add fix (80e0e5ee) also in, the renderer built from source into a nightly-80-7315957c game folder | Windows 11 Pro 10.0.29704, 2048x1152 output | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | uniform snapshot RX_VK_UB_SNAPSHOT=0 (main’s path on an Arc discrete card); frame generation off, uncapped (HOS_FPS_MAX=1000) |
the starting forest (child), the spawn spot on the player character’s balcony after the opening, about 50 s standing still | fps over the last eight 5 s windows 43.6, 48.8, 41.0, 49.4, 61.8, 69.1, 69.5, 69.4; GPU time about 9.4 ms a frame; sky where the ground and the treehouse should be, white glowing blocks (run u_off3, 04:36) |
| 2026-11-14 | 47ebdc75 (the uniform snapshot on every Intel GPU, on main cf605c85) with the Battlemage float-add fix (80e0e5ee) also in, the renderer built from source into a nightly-80-7315957c game folder | Windows 11 Pro 10.0.29704, 2048x1152 output | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | uniform snapshot on (the default with the change); frame generation off, uncapped (HOS_FPS_MAX=1000) |
the starting forest (child), the spawn spot on the player character’s balcony after the opening, about 50 s standing still | fps over the last eight 5 s windows 78.5, 79.1, 78.9, 79.5, 79.1, 79.8, 78.7, 79.0; GPU time about 6.9 ms a frame; the ground and the player character’s treehouse drawn (run u_on3, 04:34) |
| 2026-11-14 | 47ebdc75 (the uniform snapshot on every Intel GPU, on main cf605c85) with the Battlemage float-add fix (80e0e5ee) also in, the renderer built from source into a nightly-80-7315957c game folder | Windows 11 Pro 10.0.29704, 2048x1152 output | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | uniform snapshot RX_VK_UB_SNAPSHOT=0 (main’s path on an Arc discrete card); frame generation off, uncapped (HOS_FPS_MAX=1000) |
the starting forest (child), the spawn spot on the player character’s balcony after the opening, about 50 s standing still | fps over the last eight 5 s windows 68.2, 68.6, 67.2, 68.6, 68.4, 69.5, 69.0, 69.9; GPU time about 9.3 ms a frame; sky where the ground and the treehouse should be, white glowing blocks (run u_off2, 04:33) |
| 2026-11-14 | 47ebdc75 (the uniform snapshot on every Intel GPU, on main cf605c85) with the Battlemage float-add fix (80e0e5ee) also in, the renderer built from source into a nightly-80-7315957c game folder | Windows 11 Pro 10.0.29704, 2048x1152 output | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | uniform snapshot on (the default with the change); frame generation off, uncapped (HOS_FPS_MAX=1000) |
the starting forest (child), the spawn spot on the player character’s balcony after the opening, about 50 s standing still | fps over the last eight 5 s windows 78.2, 77.0, 78.8, 78.9, 78.7, 79.3, 79.3, 79.3; GPU time about 6.9 ms a frame; the ground and the player character’s treehouse drawn (run u_on2, 04:31) |
| 2026-11-14 | 47ebdc75 (the uniform snapshot on every Intel GPU, on main cf605c85) with the Battlemage float-add fix (80e0e5ee) also in, the renderer built from source into a nightly-80-7315957c game folder | Windows 11 Pro 10.0.29704, 2048x1152 output | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | uniform snapshot RX_VK_UB_SNAPSHOT=0 (main’s path on an Arc discrete card); frame generation off, uncapped (HOS_FPS_MAX=1000) |
the starting forest (child), the spawn spot on the player character’s balcony after the opening, about 50 s standing still | fps over the last eight 5 s windows 69.5, 69.7, 69.4, 69.4, 69.4, 69.6, 69.6, 69.4; GPU time about 9.4 ms a frame; sky where the ground and the treehouse should be, white glowing blocks (run u_off1, 04:29) |
| 2026-11-14 | 47ebdc75 (the uniform snapshot on every Intel GPU, on main cf605c85) with the Battlemage float-add fix (80e0e5ee) also in, the renderer built from source into a nightly-80-7315957c game folder | Windows 11 Pro 10.0.29704, 2048x1152 output | Intel Arc B580 12 GB, driver 32.0.101.9154 | Core i5-13600K, 64 GB | uniform snapshot on (the default with the change); frame generation off, uncapped (HOS_FPS_MAX=1000) |
the starting forest (child), the spawn spot on the player character’s balcony after the opening, about 50 s standing still | fps over the last eight 5 s windows 77.6, 77.9, 77.5, 77.1, 78.5, 77.9, 79.2, 79.3; GPU time about 6.9 ms a frame; the ground and the player character’s treehouse drawn (run u_on1, 04:28) |
| 2026-11-13 | longshot v3.0.0-rc.2-10-g97915751 (source d11f21e2): main 839e85ff with #155’s commits, built from source (llvm-mingw, no DLSS) | Windows 11, 1280x720 window, 2048x1152 output (16:9) | RX 9070 XT, AMD 26.10.2, Vulkan driver 0x8001a0 | Ryzen 7 5700X3D, 64 GB | HOS_NATIVE_ZSTD_FRAMES=1 |
the main outdoor area, the standard test: the save loaded, the camera turning at full speed, frame rate unlocked (HOS_FPS_MAX=240), 20 s window; one binary, the switch unset and set in 3 alternating pairs; pair 3 of 3, second of the pair |
run ls155hyrule-frames-2: fps 65.45, CPU 61.88 ms a frame, p99 21.6 ms, worst 72.4 ms, game clock 1.000; no frame over 100 ms |
| 2026-11-13 | longshot v3.0.0-rc.2-10-g97915751 (source d11f21e2): main 839e85ff with #155’s commits, built from source (llvm-mingw, no DLSS) | Windows 11, 1280x720 window, 2048x1152 output (16:9) | RX 9070 XT, AMD 26.10.2, Vulkan driver 0x8001a0 | Ryzen 7 5700X3D, 64 GB | HOS_NATIVE_ZSTD_FRAMES=1 |
the main outdoor area, the standard test: the save loaded, the camera turning at full speed, frame rate unlocked (HOS_FPS_MAX=240), 20 s window; one binary, the switch unset and set in 3 alternating pairs; pair 2 of 3, first of the pair |
run ls155hyrule-frames-1: fps 65.09, CPU 63.41 ms a frame, p99 21.0 ms, worst 72.3 ms, game clock 1.000; no frame over 100 ms |
| 2026-11-13 | longshot v3.0.0-rc.2-10-g97915751 (source d11f21e2): main 839e85ff with #155’s commits, built from source (llvm-mingw, no DLSS) | Windows 11, 1280x720 window, 2048x1152 output (16:9) | RX 9070 XT, AMD 26.10.2, Vulkan driver 0x8001a0 | Ryzen 7 5700X3D, 64 GB | HOS_NATIVE_ZSTD_FRAMES=1 |
the main outdoor area, the standard test: the save loaded, the camera turning at full speed, frame rate unlocked (HOS_FPS_MAX=240), 20 s window; one binary, the switch unset and set in 3 alternating pairs; pair 1 of 3, second of the pair |
run ls155hyrule-frames-0: fps 62.94, CPU 65.84 ms a frame, p99 24.5 ms, worst 605.1 ms, game clock 1.000; 3 frames over 100 ms (605, 128 and 114 ms) just after the window opened: the load was seen at 26.3 s, not 28.4-30.3 s as in the other five runs, so the frames after the load that every run has fell inside the window; rejected by the window rule (a frame over 250 ms), reported as data |
| 2026-11-13 | longshot v3.0.0-rc.2-10-g97915751 (source d11f21e2): main 839e85ff with #155’s commits, built from source (llvm-mingw, no DLSS) | Windows 11, 1280x720 window, 2048x1152 output (16:9) | RX 9070 XT, AMD 26.10.2, Vulkan driver 0x8001a0 | Ryzen 7 5700X3D, 64 GB | HOS_NATIVE_ZSTD_FRAMES unset |
the main outdoor area, the standard test: the save loaded, the camera turning at full speed, frame rate unlocked (HOS_FPS_MAX=240), 20 s window; one binary, the switch unset and set in 3 alternating pairs; pair 3 of 3, first of the pair |
run ls155hyrule-ctl-2: fps 66.04, CPU 61.84 ms a frame, p99 20.0 ms, worst 53.6 ms, game clock 1.001; no frame over 100 ms |
| 2026-11-13 | longshot v3.0.0-rc.2-10-g97915751 (source d11f21e2): main 839e85ff with #155’s commits, built from source (llvm-mingw, no DLSS) | Windows 11, 1280x720 window, 2048x1152 output (16:9) | RX 9070 XT, AMD 26.10.2, Vulkan driver 0x8001a0 | Ryzen 7 5700X3D, 64 GB | HOS_NATIVE_ZSTD_FRAMES unset |
the main outdoor area, the standard test: the save loaded, the camera turning at full speed, frame rate unlocked (HOS_FPS_MAX=240), 20 s window; one binary, the switch unset and set in 3 alternating pairs; pair 2 of 3, second of the pair |
run ls155hyrule-ctl-1: fps 65.99, CPU 60.56 ms a frame, p99 21.5 ms, worst 46.8 ms, game clock 1.000; no frame over 100 ms |
| 2026-11-13 | longshot v3.0.0-rc.2-10-g97915751 (source d11f21e2): main 839e85ff with #155’s commits, built from source (llvm-mingw, no DLSS) | Windows 11, 1280x720 window, 2048x1152 output (16:9) | RX 9070 XT, AMD 26.10.2, Vulkan driver 0x8001a0 | Ryzen 7 5700X3D, 64 GB | HOS_NATIVE_ZSTD_FRAMES unset |
the main outdoor area, the standard test: the save loaded, the camera turning at full speed, frame rate unlocked (HOS_FPS_MAX=240), 20 s window; one binary, the switch unset and set in 3 alternating pairs; pair 1 of 3, first of the pair |
run ls155hyrule-ctl-0: fps 63.39, CPU 63.34 ms a frame, p99 24.0 ms, worst 59.2 ms, game clock 1.001; no frame over 100 ms |
| 2026-11-13 | longshot v3.0.0-rc.2-1-g839e85ff (source ef9b723c), its renderer with #162’s commits (9ac9b714), built from source (no DLSS) | Windows 11, 1280x720 window, 2048x1152 output (16:9) | RX 9070 XT, AMD 26.10.2, Vulkan driver 0x8001a0 | Ryzen 7 5700X3D, 64 GB | RX_VK_RECORDER=thread |
the main outdoor area, the standard test: the save loaded, the camera turning at full speed, frame rate unlocked (HOS_FPS_MAX=240), 20 s window; 3 alternating pairs of four arms, iteration 1 rejected (a thread run’s clock 0.908 over 407 and 486 ms frames) |
2 runs (iterations 0 and 2 of the A/B): fps 75.54, 74.54, CPU 67.33, 69.96 ms a frame; the replay waited for the recorder 15.0 ms a second, 7.65 of them at vkResetDescriptorPool |
| 2026-11-13 | longshot v3.0.0-rc.2-1-g839e85ff (source ef9b723c), its renderer with #162’s commits (9ac9b714), built from source (no DLSS) | Windows 11, 1280x720 window, 2048x1152 output (16:9) | RX 9070 XT, AMD 26.10.2, Vulkan driver 0x8001a0 | Ryzen 7 5700X3D, 64 GB | RX_VK_RECORDER=thread, RX_VK_RECORDER_OWN=1 |
the main outdoor area, the standard test: the save loaded, the camera turning at full speed, frame rate unlocked (HOS_FPS_MAX=240), 20 s window; 3 alternating pairs of four arms, iteration 1 rejected (a thread run’s clock 0.908 over 407 and 486 ms frames) |
2 runs (iterations 0 and 2 of the A/B): fps 76.79, 76.34, CPU 64.78, 65.04 ms a frame; the replay waited for the recorder 9.3 ms a second, none at vkResetDescriptorPool |
| 2026-11-13 | longshot v3.0.0-rc.2-1-g839e85ff (source ef9b723c), its renderer with #162’s commits (9ac9b714), built from source (no DLSS) | Windows 11, 1280x720 window, 2048x1152 output (16:9) | RX 9070 XT, AMD 26.10.2, Vulkan driver 0x8001a0 | Ryzen 7 5700X3D, 64 GB | The recorder off | the main outdoor area, the standard test: the save loaded, the camera turning at full speed, frame rate unlocked (HOS_FPS_MAX=240), 20 s window; 3 alternating pairs of four arms, iteration 1 rejected (a thread run’s clock 0.908 over 407 and 486 ms frames) |
2 runs (iterations 0 and 2 of the A/B): fps 66.59, 66.44, CPU 57.95, 60.20 ms a frame |
| 2026-11-13 | longshot v3.0.0-rc.2-1-g839e85ff (source ef9b723c), its renderer with #162’s commits (9ac9b714), built from source (no DLSS) | Windows 11, 1280x720 window, 2048x1152 output (16:9) | RX 9070 XT, AMD 26.10.2, Vulkan driver 0x8001a0 | Ryzen 7 5700X3D, 64 GB | RX_VK_RECORDER=thread, RX_VK_RECORDER_IDLE=wait |
the main outdoor area, the standard test: the save loaded, the camera turning at full speed, frame rate unlocked (HOS_FPS_MAX=240), 20 s window; 3 alternating pairs of four arms, iteration 1 rejected (a thread run’s clock 0.908 over 407 and 486 ms frames) |
2 runs (iterations 0 and 2 of the A/B): fps 74.79, 73.44, CPU 66.12, 67.52 ms a frame; the replay waited for the recorder 22.1 ms a second; the recorder spun 0.32 of a core idle against 0.44 |
| 2026-11-13 | main 514591a3 | Linux (CachyOS, kernel 7.3.3), KDE Wayland | RTX 3060 Laptop 6 GB, 620.70.06 | Ryzen 7 6800H, 15 GB | A packaged builder (the AppImage’s AppRun stand-in: LONGSHOT_TOOLS set), --update of a built folder, nothing changed, --no-prepare-shaders |
The finish stage of 2 builds interleaved with perf/prompt-cache-packaged’s, after a warm-up build each (the builder’s where the time went lines); main’s runs |
prompts: regenerated (packaged or direct build) every build: finish 148 s and 191 s |
| 2026-11-13 | perf/prompt-cache-packaged b7faba3f (main 514591a3 + keeping the prompts in packaged builds) | Linux (CachyOS, kernel 7.3.3), KDE Wayland | RTX 3060 Laptop 6 GB, 620.70.06 | Ryzen 7 6800H, 15 GB | A packaged builder (the AppImage’s AppRun stand-in: LONGSHOT_TOOLS set), --update of a built folder, nothing changed, --no-prepare-shaders |
The finish stage of 2 builds interleaved with main 514591a3’s, after a warm-up build each (the builder’s where the time went lines); the branch’s runs |
prompts: unchanged, kept: finish 12 s and 11 s |
| 2026-11-13 | main 5a8ae7e3 (longshot abc-base), a full Linux build (Ubuntu 22.04 container, then the AppImage’s --cli --fsr); the same build in both arms |
Fedora Linux 45 (kernel 7.3.3), KDE Plasma (Wayland), 1280x720 window, 2560x1440 output (16:9) | RX 9070 XT 16 GB, Mesa RADV 26.2.6 (GFX1201) | Ryzen 5 5600X (6 cores, 12 threads), 31 GB | the recorder off (the default); AMD FSR 4 Native AA (fsr4-vk 4.1.1 FP8), HDR Expanded, sharpness 50%, anisotropic 16x, VRAM budget 11 GB, frame generation off; HOS_FPS_MAX=240 (uncapped), HOS_PRESENT_TRACE on |
Market (child) from one save copied fresh each run (longshot_ctl.py launch --save --slot), standing still with no input: 120 s measured after the load and 10 s; four runs interleaved off, thread, thread, off, the build’s caches warm from earlier runs |
Run 4 of 4 (21:03 local, recorder off): 41.7 fps, 1% low 32.1 fps, frame time median 23.79 ms and p99 29.40 ms (present trace); replay submit 18.75 ms a frame, 7.11 ms per 1000 draws, 2630 draws a frame, GPU 12.8 ms a frame; no device lost |
| 2026-11-13 | main 5a8ae7e3 (longshot abc-base), a full Linux build (Ubuntu 22.04 container, then the AppImage’s --cli --fsr); the same build in both arms |
Fedora Linux 45 (kernel 7.3.3), KDE Plasma (Wayland), 1280x720 window, 2560x1440 output (16:9) | RX 9070 XT 16 GB, Mesa RADV 26.2.6 (GFX1201) | Ryzen 5 5600X (6 cores, 12 threads), 31 GB | RX_VK_RECORDER=thread; AMD FSR 4 Native AA (fsr4-vk 4.1.1 FP8), HDR Expanded, sharpness 50%, anisotropic 16x, VRAM budget 11 GB, frame generation off; HOS_FPS_MAX=240 (uncapped), HOS_PRESENT_TRACE on |
Market (child) from one save copied fresh each run (longshot_ctl.py launch --save --slot), standing still with no input: 120 s measured after the load and 10 s; four runs interleaved off, thread, thread, off, the build’s caches warm from earlier runs |
Run 3 of 4 (21:00 local, recorder thread): 44.6 fps, 1% low 33.7 fps, frame time median 22.28 ms and p99 28.25 ms (present trace); replay submit 13.91 ms a frame, 5.27 ms per 1000 draws, 2641 draws a frame, GPU 12.1 ms a frame; no device lost |
| 2026-11-13 | main 5a8ae7e3 (longshot abc-base), a full Linux build (Ubuntu 22.04 container, then the AppImage’s --cli --fsr); the same build in both arms |
Fedora Linux 45 (kernel 7.3.3), KDE Plasma (Wayland), 1280x720 window, 2560x1440 output (16:9) | RX 9070 XT 16 GB, Mesa RADV 26.2.6 (GFX1201) | Ryzen 5 5600X (6 cores, 12 threads), 31 GB | RX_VK_RECORDER=thread; AMD FSR 4 Native AA (fsr4-vk 4.1.1 FP8), HDR Expanded, sharpness 50%, anisotropic 16x, VRAM budget 11 GB, frame generation off; HOS_FPS_MAX=240 (uncapped), HOS_PRESENT_TRACE on |
Market (child) from one save copied fresh each run (longshot_ctl.py launch --save --slot), standing still with no input: 120 s measured after the load and 10 s; four runs interleaved off, thread, thread, off, the build’s caches warm from earlier runs |
Run 2 of 4 (20:57 local, recorder thread): 44.5 fps, 1% low 33.0 fps, frame time median 22.33 ms and p99 28.91 ms (present trace); replay submit 13.73 ms a frame, 5.21 ms per 1000 draws, 2636 draws a frame, GPU 12.1 ms a frame; no device lost |
| 2026-11-13 | main 5a8ae7e3 (longshot abc-base), a full Linux build (Ubuntu 22.04 container, then the AppImage’s --cli --fsr); the same build in both arms |
Fedora Linux 45 (kernel 7.3.3), KDE Plasma (Wayland), 1280x720 window, 2560x1440 output (16:9) | RX 9070 XT 16 GB, Mesa RADV 26.2.6 (GFX1201) | Ryzen 5 5600X (6 cores, 12 threads), 31 GB | the recorder off (the default); AMD FSR 4 Native AA (fsr4-vk 4.1.1 FP8), HDR Expanded, sharpness 50%, anisotropic 16x, VRAM budget 11 GB, frame generation off; HOS_FPS_MAX=240 (uncapped), HOS_PRESENT_TRACE on |
Market (child) from one save copied fresh each run (longshot_ctl.py launch --save --slot), standing still with no input: 120 s measured after the load and 10 s; four runs interleaved off, thread, thread, off, the build’s caches warm from earlier runs |
Run 1 of 4 (20:54 local, recorder off): 41.9 fps, 1% low 32.7 fps, frame time median 23.78 ms and p99 28.99 ms (present trace); replay submit 18.64 ms a frame, 7.02 ms per 1000 draws, 2636 draws a frame, GPU 12.7 ms a frame; no device lost |
| 2026-11-13 | brett-d/legendary-broccoli 05f4c4aa cherry-picked on Longshot main fbd42024 with renderer-only changes; longshot_vk.so built alone on a host of 000f19fb; one binary, the PR’s switches alternated |
SteamOS 3.9.5 (Steam Deck OLED), headless | Steam Deck OLED GPU (RADV Mesa 26.2.0-devel), GPU clock fixed at 1600 MHz | Steam Deck OLED APU, 16 GB shared RAM | Minimum preset, RX_VK_LOAD=create; 4 settings: both parts off (RX_VK_KEEP_MODULES=1, both reserves 0), modules kept (RX_VK_KEEP_MODULES=1), pressure rule off (both reserves 0), and as it is |
A save standing still on a heavy view (about 2,150 draws a frame), 720 presents a run, 3 rounds; and one pair in the starting forest (p0lux A/B) | Medians: RssAnon 2,661 / 2,566 / 1,696 / 1,718 MB, MemAvailable 1,737 / 1,699 / 2,586 / 2,707 MB, fps 30.82 / 30.92 / 30.75 / 30.85 (both off / modules kept / pressure off / as it is); 4643 shader modules let go after the precompile. starting-area (both off, then as it is): RssAnon 2,678 -> 1,710 MB, MemAvailable 1,354 -> 2,312 MB, 27.37 and 27.33 fps |
| 2026-11-13 | brett-d/legendary-broccoli 68d00063 (Longshot main 514591a3 + this change), host and renderer built from source | SteamOS 3.9.5, kernel 7.2.7-valve2 (Steam Deck OLED, Gaming Mode 1280x720) | AMD Custom GPU 0932 (RADV Mesa 26.2.0-devel), mesh shaders on compute fallback | Steam Deck OLED APU (stock clocks), 16 GB shared RAM, 8 GB swap file, memory guard (< 500 MB avail for 5 s with < 4 GB swap free) | Quality preset Auto (this PC), in its own folder and caches, without the Steam Deck profile, warm start (47,797 program keys kept) | the starting forest moving route (warm run 3 of #223 at 68d00063) | This change (68d00063, 1 warm run): completed the route (0 kills; the guard dipped 8 times, each with over 4 GB swap free); real fps 12.8, shown 23.1, p95 100 ms; start-up 9.3 s, 38,180 shader modules let go after the precompile; peak RSS 9,083 MB, peak anon 5,968 MB, process swap 1,237 MB, MemAvailable minimum 324 MB (6,954 MB swap free then); 3 vk: memory pressure on Linux: lines (textures’ budget 1,561 -> 1,590 -> 2,054 MB) |
| 2026-11-13 | brett-d/legendary-broccoli 05f4c4aa (Longshot main 7315957c + this change), against Longshot main 08771847, host and renderer built from source | SteamOS 3.9.5, kernel 7.2.7-valve2 (Steam Deck OLED, Gaming Mode 1280x720) | AMD Custom GPU 0932 (RADV Mesa 26.2.0-devel), mesh shaders on compute fallback | Steam Deck OLED APU (stock clocks), 16 GB shared RAM, 8 GB swap file, memory guard (< 500 MB avail for 5 s with < 4 GB swap free) | Quality preset Auto (this PC), each build in its own folder and caches, warm start (47,797 program keys kept) | the starting forest moving route (warm runs 1 and 2 of 05f4c4aa vs warm runs 1 and 2 of main 08771847) | This change (05f4c4aa, 2 warm runs): completed both runs (0 kills); real fps 14.8 and 13.8 (median 14.3), shown fps 27.3 and 26.1 (median 26.7), p95 90.3 and 91.5 ms (median 90.9 ms); start-up 8.2 and 10.7 s, 41,044 and 41,052 shader modules let go after the precompile; in-play RssAnon 0.5-1.4 GB (start-up peak 2,950 and 9,194 MB before module drop), peak RSS 8,094 and 10,173 MB, process swap 947 and 938 MB (median 942.5 MB), MemAvailable minimum 284 and 695 MB (7,192 and 5,695 MB swap free then); 3 vk: memory pressure on Linux: lines each (run 1: 2,371 -> 1,465 -> 1,465 MB; run 2: 2,326 -> 1,943 -> 1,740 MB). Main (08771847, 2 warm runs): killed by the memory guard at 2 min in both runs; peak RSS 10,738 and 10,742 MB (median 10,740 MB), peak anon 8,888 and 8,902 MB (median 8,895 MB), process swap 2,818 and 3,475 MB (median 3,146.5 MB), MemAvailable minimum 390 and 354 MB |
| 2026-11-13 | brett-d/legendary-broccoli 68d00063 (Longshot main 514591a3 + this change), host and renderer built from source | SteamOS 3.9.5, kernel 7.2.7-valve2 (Steam Deck OLED, Gaming Mode 1280x720) | AMD Custom GPU 0932 (RADV Mesa 26.2.0-devel), mesh shaders on compute fallback | Steam Deck OLED APU (stock clocks), 16 GB shared RAM, 8 GB swap file, memory guard (< 500 MB avail for 5 s with < 4 GB swap free) | Quality preset Auto (this PC), in its own folder and caches, without the Steam Deck profile, cold start | the starting forest moving route (cold run 2 of #223 at 68d00063) | This change (68d00063, 1 cold run): completed the route (0 kills); real fps 14.8, shown 27.0, p95 94 ms; 22,402 shader modules let go after the precompile; peak RSS 6,954 MB, peak anon 5,697 MB, process swap 3,031 MB, MemAvailable minimum 508 MB (3,609 MB swap free then); 3 vk: memory pressure on Linux: lines |
| 2026-11-13 | brett-d/legendary-broccoli 05f4c4aa (Longshot main 7315957c + this change), against Longshot main 08771847, host and renderer built from source | SteamOS 3.9.5, kernel 7.2.7-valve2 (Steam Deck OLED, Gaming Mode 1280x720) | AMD Custom GPU 0932 (RADV Mesa 26.2.0-devel), mesh shaders on compute fallback | Steam Deck OLED APU (stock clocks), 16 GB shared RAM, 8 GB swap file, memory guard (< 500 MB avail for 5 s with < 4 GB swap free) | Quality preset Auto (this PC), each build in its own folder and caches, cold start | the starting forest moving route (cold run 1 of 05f4c4aa vs cold run 1 of main 08771847) | This change (05f4c4aa, 1 cold run): completed the route, 14.4 real fps (27.0 shown), p95 96 ms; 25,410 shader modules let go after the precompile; peak RSS 8,979 MB, peak anon 8,003 MB, process swap 2,892 MB, MemAvailable minimum 610 MB (4.7 GB swap free then; no kill); 4 vk: memory pressure on Linux: lines (textures’ budget held at 1,732 MB x3, then recovered to 2,560 MB once memory returned). Main (08771847, 1 cold run): killed by the memory guard at 6 min (14.0 real fps before the kill); peak RSS 9,371 MB, peak anon 6,069 MB, process swap 4,598 MB, MemAvailable minimum 387 MB |
| 2026-11-13 | game v3.0.0-rc.3-26-g3fb77017 (fix/input-script-background 3fb77017) | Windows 11 Pro 26300, a window that never takes focus (-Window) |
RTX 3080 10 GB, 621.14 | i9-12900K, 64 GB | Default; HOS_INPUT_SCRIPT_BACKGROUND=1, HOS_INPUT_LIVE, HOS_SKIP_OPENING=1, no HOS_INPUT_EXCLUSIVE |
The test save, in play, the window behind others; the user pressed Enter in other windows and typed in a chat | 0 pauses in the log, Link didn’t move: the player’s keyboard stays out behind another window (run pr) |
| 2026-11-13 | game v3.0.0-rc.3-15-gfbd42024 (main fbd42024, host/nn_hid.c as main 514591a3) |
Windows 11 Pro 26300, a window that never takes focus (-Window) |
RTX 3080 10 GB, 621.14 | i9-12900K, 64 GB | Default; HOS_INPUT_SCRIPT_BACKGROUND=1, HOS_INPUT_LIVE, HOS_SKIP_OPENING=1, no HOS_INPUT_EXCLUSIVE |
The test save, in play, the window behind others; the user pressed Enter in other windows and typed in a chat | 3 pauses and unpauses (control: state play -> menu), Link jumped and moved while the user typed elsewhere (run main) |
| 2026-11-13 | v3.0.0-rc.3-27-g6746c9e1 (feat/hud-layout, the loading icon placed as the HUD’s, on main 514591a3, llvm-mingw, NVIDIA DLSS SDK) against v2.1.0-rc.3-330-g82de567a (main) | Windows 11 Pro (26300), headless (offscreen, scripted input through the control bridge) | RTX 5080 16 GB, 621.24 | Ryzen 9 9900X3D, 64 GB | HOS_OUTPUT_RES 1440x1080 (4:3) and 2560x1080 (21:9), Ultrawide menus Centred, HOS_WIDE_UI_LOG, no sound, HOS_NO_PACING=1 |
the starting forest walked out to the main outdoor area, a shot every 2 presents through the loading fade | The loading icon whole at the screen’s bottom-right corner at both shapes (its group moved -300, down 135 at 4:3; 640 at 21:9); on 7cedec2d it was off the screen at 4:3, on main squeezed. The fade still black over the whole picture (run03) |
| 2026-11-13 | v3.0.0-rc.3-28-g5487b3f1 (feat/hud-layout’s review round on main 514591a3, llvm-mingw, NVIDIA DLSS SDK) against v2.2.0-rc.1-29-g607a186c (the PR’s head before it) and v2.1.0-rc.3-330-g82de567a (main), one run each, on the same game files | Windows 11 Pro (26300), headless (offscreen, scripted input through the control bridge) | RTX 5080 16 GB, 621.24 | Ryzen 9 9900X3D, 64 GB | HOS_OUTPUT_RES 2560x1080 (21:9) and 1440x1080 (4:3), Ultrawide menus Centred, HOS_WIDE_UI_LOG, no sound, HOS_NO_PACING=1 |
the starting forest walked out to the main outdoor area (a shot every 2 presents through the loading fade), then 1,300 presents each in play, with the map open (View) and with the pause menu open | The loading fade over the whole picture: main at 21:9 and 607a186c at both shapes left the sides (bands) lit while the middle was black (mean 230 and 104 of 255), this head 0. The plan per present, 607a186c then this head: play 0.10-0.11 then 0.05-0.06 ms, the map 0.09-0.12 then 0.05, the pause menu 0.13-0.19 then 0.05-0.06. The map is drawn as a menu (fitted at 4:3, not moved at 21:9). No text box came up (the companion had nothing to say) (run02) |
| 2026-11-13 | v2.1.0-rc.3-337-g9c5a39c1 (feat/hud-layout on main ef8b784f, llvm-mingw, NVIDIA DLSS SDK) against v2.1.0-rc.3-330-g82de567a (main), two folders on the same game files, runs interleaved main, PR, main, PR | Windows 11 Pro (26300), headless (offscreen, scripted input through the control bridge) | RTX 5080 16 GB, 621.24 | Ryzen 9 9900X3D, 64 GB | HOS_OUTPUT_RES 3440x1440 (21:9) and 1440x1080 (4:3), the options’ DLSS row, no sound, HOS_NO_PACING=1 |
the starting forest (the debug teleport’s arrival), standing 1,800 presents | Frame interval median of the two runs: 21:9 main 11.44 ms, PR 11.37 ms; 4:3 main 11.19 ms, PR 11.21 ms (within run-to-run spread). The HUD’s plan 0.10-0.11 ms a present on the presenting thread (HOS_WIDE_UI_LOG). HUD at its 16:9 size in the corners at 5:4, 4:3, 3:2, 16:10, 21:9, 12:5; 32:9 to 12:5’s sides; 48:9 on the middle screen; 16:9 as main (run01) |
| 2026-11-13 | #231 @ 53a83cca on main e28236b4 (full make_game, shaders prepared, its own game folder, a warm-up start per folder) | Windows 11 | NVIDIA GeForce RTX 5090, driver 621.14 | Intel Core i7-13700K (16 cores, 24 threads), 96 GB | MP165: 2048x1152 output in a 1600x900 window, DLSS (dlss=1, upscaling=2, resolution=1), frame generation 2x (frame_generation=2, frame_generation_by=0; the log: V-Sync Auto with DLSS Frame Generation at 2x: FIFO; the renderer’s own frame generation made the generated presents, DLSS-G 0 of 600), frame_rate_cap=4, HOS_FPS_MAX=165 | the starting forest, the scripted moving route (~129 s, measured 30-110 s; game frames from presents.csv, 1% low = the slowest 1% of game frames as fps), interleaved main / #231 (run lsflat-mp165-r2-pr, 02:59 local) | 71.98 real fps (143.9 shown), 1% low 40.2 fps, p50 13.77 ms, p99 20.37 ms; replay submit / draw / descriptors 8.78 / 4.82 / 1.54 ms a frame |
| 2026-11-13 | main e28236b4 (full make_game, shaders prepared, its own game folder, a warm-up start per folder) | Windows 11 | NVIDIA GeForce RTX 5090, driver 621.14 | Intel Core i7-13700K (16 cores, 24 threads), 96 GB | MP165: 2048x1152 output in a 1600x900 window, DLSS (dlss=1, upscaling=2, resolution=1), frame generation 2x (frame_generation=2, frame_generation_by=0; the log: V-Sync Auto with DLSS Frame Generation at 2x: FIFO; the renderer’s own frame generation made the generated presents, DLSS-G 0 of 600), frame_rate_cap=4, HOS_FPS_MAX=165 | the starting forest, the scripted moving route (~129 s, measured 30-110 s; game frames from presents.csv, 1% low = the slowest 1% of game frames as fps), interleaved main / #231 (run lsflat-mp165-r2-main, 02:56 local) | 69.99 real fps (140.1 shown), 1% low 38.5 fps, p50 14.03 ms, p99 22.14 ms; replay submit / draw / descriptors 9.22 / 5.19 / 1.66 ms a frame |
| 2026-11-13 | #231 @ 53a83cca on main e28236b4 (full make_game, shaders prepared, its own game folder, a warm-up start per folder) | Windows 11 | NVIDIA GeForce RTX 5090, driver 621.14 | Intel Core i7-13700K (16 cores, 24 threads), 96 GB | MP165: 2048x1152 output in a 1600x900 window, DLSS (dlss=1, upscaling=2, resolution=1), frame generation 2x (frame_generation=2, frame_generation_by=0; the log: V-Sync Auto with DLSS Frame Generation at 2x: FIFO; the renderer’s own frame generation made the generated presents, DLSS-G 0 of 600), frame_rate_cap=4, HOS_FPS_MAX=165 | the starting forest, the scripted moving route (~129 s, measured 30-110 s; game frames from presents.csv, 1% low = the slowest 1% of game frames as fps), interleaved main / #231 (run lsflat-mp165-r1-pr, 02:53 local) | EXCLUDED with its pair (main’s run had outside CPU load): 71.08 real fps (142.3 shown), 1% low 42.8 fps, p50 13.85 ms, p99 20.99 ms; replay submit / draw / descriptors 8.88 / 4.92 / 1.57 ms a frame |
| 2026-11-13 | main e28236b4 (full make_game, shaders prepared, its own game folder, a warm-up start per folder) | Windows 11 | NVIDIA GeForce RTX 5090, driver 621.14 | Intel Core i7-13700K (16 cores, 24 threads), 96 GB | MP165: 2048x1152 output in a 1600x900 window, DLSS (dlss=1, upscaling=2, resolution=1), frame generation 2x (frame_generation=2, frame_generation_by=0; the log: V-Sync Auto with DLSS Frame Generation at 2x: FIFO; the renderer’s own frame generation made the generated presents, DLSS-G 0 of 600), frame_rate_cap=4, HOS_FPS_MAX=165 | the starting forest, the scripted moving route (~129 s, measured 30-110 s; game frames from presents.csv, 1% low = the slowest 1% of game frames as fps), interleaved main / #231 (run lsflat-mp165-r1-main, 02:51 local) | EXCLUDED (outside CPU load: process cores 4.54 against ~5.9, 132 frames over 50 ms, ~22 fps for ~20 s): 57.46 real fps (115.0 shown), 1% low 13.9 fps, p50 14.27 ms, p99 62.16 ms; replay submit / draw / descriptors 10.35 / 5.68 / 1.81 ms a frame |
| 2026-11-13 | #231 @ 53a83cca on main e28236b4 (full make_game, shaders prepared, its own game folder, a warm-up start per folder) | Windows 11 | NVIDIA GeForce RTX 5090, driver 621.14 | Intel Core i7-13700K (16 cores, 24 threads), 96 GB | MP165: 2048x1152 output in a 1600x900 window, DLSS (dlss=1, upscaling=2, resolution=1), frame generation 2x (frame_generation=2, frame_generation_by=0; the log: V-Sync Auto with DLSS Frame Generation at 2x: FIFO; the renderer’s own frame generation made the generated presents, DLSS-G 0 of 600), frame_rate_cap=4, HOS_FPS_MAX=165 | the starting forest, the scripted moving route (~129 s, measured 30-110 s; game frames from presents.csv, 1% low = the slowest 1% of game frames as fps), interleaved main / #231 (run lsflat-mp165-r0-pr, 02:42 local) | 72.23 real fps (144.4 shown), 1% low 40.4 fps, p50 13.69 ms, p99 21.66 ms; replay submit / draw / descriptors 8.79 / 4.84 / 1.54 ms a frame |
| 2026-11-13 | main e28236b4 (full make_game, shaders prepared, its own game folder, a warm-up start per folder) | Windows 11 | NVIDIA GeForce RTX 5090, driver 621.14 | Intel Core i7-13700K (16 cores, 24 threads), 96 GB | MP165: 2048x1152 output in a 1600x900 window, DLSS (dlss=1, upscaling=2, resolution=1), frame generation 2x (frame_generation=2, frame_generation_by=0; the log: V-Sync Auto with DLSS Frame Generation at 2x: FIFO; the renderer’s own frame generation made the generated presents, DLSS-G 0 of 600), frame_rate_cap=4, HOS_FPS_MAX=165 | the starting forest, the scripted moving route (~129 s, measured 30-110 s; game frames from presents.csv, 1% low = the slowest 1% of game frames as fps), interleaved main / #231 (run lsflat-mp165-r0-main, 02:39 local) | 67.49 real fps (135.1 shown), 1% low 35.0 fps, p50 14.53 ms, p99 23.46 ms; replay submit / draw / descriptors 9.51 / 5.32 / 1.69 ms a frame |
| 2026-11-13 | #231 @ 53a83cca on main e28236b4 (full make_game, shaders prepared, its own game folder, a warm-up start per folder) | Windows 11 | NVIDIA GeForce RTX 5090, driver 621.14 | Intel Core i7-13700K (16 cores, 24 threads), 96 GB | M1152: 2048x1152 output in a 1600x900 window, mailbox, frame generation off, no DLSS (options dlss=0, frame_generation=0, frame_rate_cap=5), HOS_FPS_MAX=240 | the starting forest, the scripted moving route (~129 s, measured 30-110 s; game frames from presents.csv, 1% low = the slowest 1% of game frames as fps), interleaved main / #231 (run lsflat-m1152-r1-pr, 02:46 local) | 77.00 fps, 1% low 45.3 fps, p50 12.89 ms, p99 18.23 ms; replay submit / draw / descriptors 8.16 / 4.69 / 1.53 ms a frame |
| 2026-11-13 | main e28236b4 (full make_game, shaders prepared, its own game folder, a warm-up start per folder) | Windows 11 | NVIDIA GeForce RTX 5090, driver 621.14 | Intel Core i7-13700K (16 cores, 24 threads), 96 GB | M1152: 2048x1152 output in a 1600x900 window, mailbox, frame generation off, no DLSS (options dlss=0, frame_generation=0, frame_rate_cap=5), HOS_FPS_MAX=240 | the starting forest, the scripted moving route (~129 s, measured 30-110 s; game frames from presents.csv, 1% low = the slowest 1% of game frames as fps), interleaved main / #231 (run lsflat-m1152-r1-main, 02:44 local) | 75.75 fps, 1% low 43.8 fps, p50 13.07 ms, p99 19.04 ms; replay submit / draw / descriptors 8.45 / 4.88 / 1.60 ms a frame |
| 2026-11-13 | #231 @ 53a83cca on main e28236b4 (full make_game, shaders prepared, its own game folder, a warm-up start per folder) | Windows 11 | NVIDIA GeForce RTX 5090, driver 621.14 | Intel Core i7-13700K (16 cores, 24 threads), 96 GB | M1152: 2048x1152 output in a 1600x900 window, mailbox, frame generation off, no DLSS (options dlss=0, frame_generation=0, frame_rate_cap=5), HOS_FPS_MAX=240 | the starting forest, the scripted moving route (~129 s, measured 30-110 s; game frames from presents.csv, 1% low = the slowest 1% of game frames as fps), interleaved main / #231 (run lsflat-m1152-r0-pr, 02:37 local) | 73.71 fps, 1% low 40.7 fps, p50 13.38 ms, p99 20.36 ms; replay submit / draw / descriptors 8.40 / 4.75 / 1.56 ms a frame |
| 2026-11-13 | main e28236b4 (full make_game, shaders prepared, its own game folder, a warm-up start per folder) | Windows 11 | NVIDIA GeForce RTX 5090, driver 621.14 | Intel Core i7-13700K (16 cores, 24 threads), 96 GB | M1152: 2048x1152 output in a 1600x900 window, mailbox, frame generation off, no DLSS (options dlss=0, frame_generation=0, frame_rate_cap=5), HOS_FPS_MAX=240 | the starting forest, the scripted moving route (~129 s, measured 30-110 s; game frames from presents.csv, 1% low = the slowest 1% of game frames as fps), interleaved main / #231 (run lsflat-m1152-r0-main, 02:35 local) | 74.20 fps, 1% low 40.5 fps, p50 13.32 ms, p99 20.42 ms; replay submit / draw / descriptors 8.63 / 5.00 / 1.65 ms a frame |
| 2026-11-13 | #231 @ 559f480 merged on main d964886 + #225 (e298bd4), the renderer swapped into a copy of the user’s folder | SteamOS 3.9.5 (kernel 7.2.7-valve2), docked to 3440x1440 (gamescope) | Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel | AMD Custom APU 0932 (4 cores, 8 threads), 16 GB (RSS 7.5-8.3 GB, swapping) | The user’s options (Deck profile: output 2048x1152, FSR 3.1 ~50%, frame generation 2x), stock clocks, memory guard on | the starting forest, the moving route (125 s, measured 30-110 s), order base, #231, #232, #156, #156, #232, #231, base (run X231-2) | 14.7 real fps (26.5 shown), p95 108 ms, 123 frames over 100 ms; replay submit / draw / descriptors 19.2 / 10.8 / 6.24 ms a frame; 870k pages swapped out |
| 2026-11-13 | #231 @ 559f480 merged on main d964886 + #225 (e298bd4), the renderer swapped into a copy of the user’s folder | SteamOS 3.9.5 (kernel 7.2.7-valve2), docked to 3440x1440 (gamescope) | Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel | AMD Custom APU 0932 (4 cores, 8 threads), 16 GB (RSS 7.5-8.3 GB, swapping) | The user’s options (Deck profile: output 2048x1152, FSR 3.1 ~50%, frame generation 2x), stock clocks, memory guard on | the starting forest, the moving route (125 s, measured 30-110 s), order base, #231, #232, #156, #156, #232, #231, base (run X231-1) | 14.1 real fps (26.0 shown), p95 101 ms, 109 frames over 100 ms; replay submit / draw / descriptors 22.0 / 12.8 / 7.01 ms a frame; 995k pages swapped out |
| 2026-11-13 | main d964886 + #225 (e298bd4), the renderer swapped into a copy of the user’s folder | SteamOS 3.9.5 (kernel 7.2.7-valve2), docked to 3440x1440 (gamescope) | Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel | AMD Custom APU 0932 (4 cores, 8 threads), 16 GB (RSS 7.5-8.3 GB, swapping) | The user’s options (Deck profile: output 2048x1152, FSR 3.1 ~50%, frame generation 2x), stock clocks, memory guard on | the starting forest, the moving route (125 s, measured 30-110 s), order base, #231, #232, #156, #156, #232, #231, base (run LB-x2) | 14.7 real fps (27.6 shown), p95 84 ms, 72 frames over 100 ms; replay submit / draw / descriptors 23.8 / 13.8 / 6.43 ms a frame; 390k pages swapped out |
| 2026-11-13 | main d964886 + #225 (e298bd4), the renderer swapped into a copy of the user’s folder | SteamOS 3.9.5 (kernel 7.2.7-valve2), docked to 3440x1440 (gamescope) | Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel | AMD Custom APU 0932 (4 cores, 8 threads), 16 GB (RSS 7.5-8.3 GB, swapping) | The user’s options (Deck profile: output 2048x1152, FSR 3.1 ~50%, frame generation 2x), stock clocks, memory guard on | the starting forest, the moving route (125 s, measured 30-110 s), order base, #231, #232, #156, #156, #232, #231, base (run LB-x1) | 15.8 real fps (30.2 shown), p95 79 ms, 74 frames over 100 ms; replay submit / draw / descriptors 20.1 / 11.9 / 6.46 ms a frame; 673k pages swapped out |
| 2026-11-13 | v2.1.0-rc.3-413-g514591a3 (source 2af40fe2); host main 514591a3, renderer b319780c (renderer-only candidate pack); upper83-pr | Windows 11 Home 26300, 3440x1440 borderless client | RTX 4070 12 GB / NVIDIA 621.01 | i7-13700F / 32 GB | 3440x1440 output, 83%, K/L/M, sharpness 0, FG/Reflex/V-Sync off, uncapped; Native HDR, Vanilla, white 200/peak 400 nits | Market, child, fountain standing view from the same copied local save | 12 s settle then >=25 s and >=1,200 presents per mode; quiet checks and continuous size guard passed. K83: 19.288 ms, GPU 11.20 ms, 114.07 W; L83: 22.813 ms, GPU 15.20 ms, 134.85 W; M83: 21.304 ms, GPU 13.70 ms, 131.03 W. Control/validation only, one run per arm; no speedup claim. |
| 2026-11-13 | v2.1.0-rc.3-413-g514591a3 (source 2af40fe2); host main 514591a3, renderer b319780c (renderer-only candidate pack); pair3-pr-c | Windows 11 Home 26300, 3440x1440 borderless client | RTX 4070 12 GB / NVIDIA 621.01 | i7-13700F / 32 GB | 3440x1440 output, 67%, K/L/M, sharpness 0, FG/Reflex/V-Sync off, uncapped; HDR Native selected but display advertised SDR; actual SDR | Market, child, fountain standing view from the same copied local save | 12 s settle then >=25 s and >=1,200 presents per mode; quiet checks and continuous size guard passed. native: 17.317 ms, GPU 10.00 ms, 104.88 W; K67: 17.143 ms, GPU 9.90 ms, 103.88 W; L67: 18.727 ms, GPU 11.80 ms, 120.48 W; M67: 17.699 ms, GPU 11.00 ms, 117.16 W. Part of main/PR/PR/main sequence; report median of two runs each, not this arm alone. |
| 2026-11-13 | v2.1.0-rc.3-413-g514591a3 (source 2af40fe2); host main 514591a3, renderer b319780c (renderer-only candidate pack); pair3-pr-b | Windows 11 Home 26300, 3440x1440 borderless client | RTX 4070 12 GB / NVIDIA 621.01 | i7-13700F / 32 GB | 3440x1440 output, 67%, K/L/M, sharpness 0, FG/Reflex/V-Sync off, uncapped; HDR Native selected but display advertised SDR; actual SDR | Market, child, fountain standing view from the same copied local save | 12 s settle then >=25 s and >=1,200 presents per mode; quiet checks and continuous size guard passed. M67: 18.809 ms, GPU 11.40 ms, 113.32 W; L67: 19.165 ms, GPU 11.90 ms, 118.40 W; K67: 16.899 ms, GPU 9.80 ms, 104.75 W; native: 16.239 ms, GPU 9.60 ms, 107.09 W. Part of main/PR/PR/main sequence; report median of two runs each, not this arm alone. |
| 2026-11-13 | v2.1.0-rc.3-413-g514591a3 (source 2af40fe2); host main 514591a3, renderer main 514591a3; pair3-main-d | Windows 11 Home 26300, 3440x1440 borderless client | RTX 4070 12 GB / NVIDIA 621.01 | i7-13700F / 32 GB | 3440x1440 output, 67%, K/L/M, sharpness 0, FG/Reflex/V-Sync off, uncapped; HDR Native selected but display advertised SDR; actual SDR | Market, child, fountain standing view from the same copied local save | 12 s settle then >=25 s and >=1,200 presents per mode; quiet checks and continuous size guard passed. M67: 21.016 ms, GPU 13.70 ms, 128.26 W; L67: 22.139 ms, GPU 15.15 ms, 137.25 W; K67: 16.787 ms, GPU 9.80 ms, 106.46 W; native: 16.125 ms, GPU 9.50 ms, 108.26 W. Part of main/PR/PR/main sequence; report median of two runs each, not this arm alone. |
| 2026-11-13 | v2.1.0-rc.3-413-g514591a3 (source 2af40fe2); host main 514591a3, renderer main 514591a3; pair3-main-a | Windows 11 Home 26300, 3440x1440 borderless client | RTX 4070 12 GB / NVIDIA 621.01 | i7-13700F / 32 GB | 3440x1440 output, 67%, K/L/M, sharpness 0, FG/Reflex/V-Sync off, uncapped; HDR Native selected but display advertised SDR; actual SDR | Market, child, fountain standing view from the same copied local save | 12 s settle then >=25 s and >=1,200 presents per mode; quiet checks and continuous size guard passed. native: 17.343 ms, GPU 9.90 ms, 102.52 W; K67: 17.241 ms, GPU 9.90 ms, 103.83 W; L67: 21.856 ms, GPU 15.00 ms, 137.14 W; M67: 20.015 ms, GPU 13.30 ms, 130.98 W. Part of main/PR/PR/main sequence; report median of two runs each, not this arm alone. |
| 2026-11-13 | v2.1.0-rc.3-413-g514591a3 (source 2af40fe2); host main 514591a3, renderer b319780c (renderer-only candidate pack); aa3-pr-b | Windows 11 Home 26300, 3440x1440 borderless client | RTX 4070 12 GB / NVIDIA 621.01 | i7-13700F / 32 GB | 3440x1440 output, Native AA, K/L/M, sharpness 0, FG/Reflex/V-Sync off, uncapped; Native HDR, Vanilla, white 200/peak 400 nits | Market, child, fountain standing view from the same copied local save | 12 s settle then >=25 s and >=1,200 presents per mode; quiet checks and continuous size guard passed. MAA: 24.250 ms, GPU 16.50 ms, 140.41 W; LAA: 25.363 ms, GPU 17.80 ms, 147.58 W; KAA: 19.720 ms, GPU 12.30 ms, 123.38 W; native: 17.018 ms, GPU 9.90 ms, 111.35 W. Control/validation only, one run per arm; no speedup claim. |
| 2026-11-13 | v2.1.0-rc.3-413-g514591a3 (source 2af40fe2); host main 514591a3, renderer main 514591a3; aa3-main-a | Windows 11 Home 26300, 3440x1440 borderless client | RTX 4070 12 GB / NVIDIA 621.01 | i7-13700F / 32 GB | 3440x1440 output, Native AA, K/L/M, sharpness 0, FG/Reflex/V-Sync off, uncapped; Native HDR, Vanilla, white 200/peak 400 nits | Market, child, fountain standing view from the same copied local save | 12 s settle then >=25 s and >=1,200 presents per mode; quiet checks and continuous size guard passed. KAA: 20.591 ms, GPU 12.30 ms, 119.26 W; LAA: 25.033 ms, GPU 17.50 ms, 146.36 W; MAA: 23.273 ms, GPU 16.10 ms, 142.57 W; native: 16.841 ms, GPU 9.70 ms, 110.81 W. Control/validation only, one run per arm; no speedup claim. |
| 2026-11-13 | host and game folder built from main 000f19fb; longshot_vk.so built alone from main fbd42024 and from the BC3 scratch’s change on it, 000cc9f2, same flags |
SteamOS 3.9.5 (kernel 7.2.7-valve2), headless | Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel; the GPU’s clock fixed at 1600 MHz (Steam’s performance menu) | AMD Custom APU 0932 (4 cores, 8 threads), 16 GB shared with the GPU | Minimum preset (RX_VK_LOW_SPEC=1), handheld profile (HOS_HANDHELD=1), the game’s own upscaler, RX_VK_VRAM_MB=2400, RX_VK_TEX_MAX_DIM=1024, RX_VK_DESC_SLOTGEN=1, RX_VK_LOAD=create, HOS_RESERVE_CORE=0 |
a save loaded by script, standing still: a heavy view (about 2,150 draws a frame), the starting forest (about 2,600), the main outdoor area (about 1,600) | p0lux run01, 7 pairs alternated (heavy view 3, the others 2), each run measured over exactly 720 presents after 3 log windows, then the driver’s count of the process’s GPU memory read from /proc/<pid>/fdinfo (drm-memory-gtt, system memory on this machine) and MemAvailable. GPU memory, main then the change: heavy view 3,884 / 3,908 / 3,800 MB against 3,675 / 3,772 / 3,737; the starting forest 4,077 / 4,064 against 3,901 / 3,919; the main outdoor area 3,482 / 3,510 against 3,362 / 3,371; the 7 pairs’ differences -209, -136, -63, -176, -145, -120, -139 MB, median -139 (the scratch is 140 MB smaller). MemAvailable: median of the 7 differences +145 MB. fps unchanged (heavy view 24.34 against 24.47, medians), start to in game 40 s against 38 s (medians of 7). The renderer’s own pictures (RX_VK_SHOTS), standing still, the change against two runs of main: mean difference of a channel (0 to 255) 8.1 / 12.0 / 5.7 against the run of main in the same state and 18.9 / 20.4 / 17.1 against the other, where main against main is 18.9 / 18.6 / 18.3. No errors in the 23 runs of the two |
| 2026-11-12 | Hookshot: v2.1.0-rc.3-249-g564ea648 and later builds of this change (sparse VRAM mirrors, https://github.com/NotRexed/Hookshot/pull/2794) on Hookshot main c544828f (unmerged then); against the last v2.1.0-rc.3 on 6dcc15c7 | Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32; nothing else running, RX_VK_TEST_ALLOWED_MB=10200; RX_VK_MIRROR_SPARSE=1 against unset (then off) |
the starting forest, standing, the camera turning | run04: on 58.4-58.7 fps, VRAM mean 7,462-7,507 MB; off 58.8 fps, 8,215 MB; screenshots alike |
| 2026-11-12 | Hookshot: v2.1.0-rc.3-249-g564ea648 and later builds of this change (sparse VRAM mirrors, https://github.com/NotRexed/Hookshot/pull/2794) on Hookshot main c544828f (unmerged then); against the last v2.1.0-rc.3 on 6dcc15c7 | Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32; nothing else running, RX_VK_TEST_ALLOWED_MB=10200; RX_VK_MIRROR_SPARSE=1 against unset (then off), and with RX_VK_SPARSE_WAIT=1 |
Castle Town’s south gate to the main outdoor area (an area change) | run03: on 55.4 / 52.5 / 54.8 fps, off 54.0 / 52.7; mirrors with memory 1,267-1,311 of 2,128-2,144 MB; the load’s longest frame 866-961 ms on, 816-902 off, 1,197-1,232 with the binds waited for on the CPU |
| 2026-11-12 | Hookshot: v2.1.0-rc.3-249-g564ea648 and later builds of this change (sparse VRAM mirrors, https://github.com/NotRexed/Hookshot/pull/2794) on Hookshot main c544828f (unmerged then); against the last v2.1.0-rc.3 on 6dcc15c7 | Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32; nothing else running, RX_VK_TEST_ALLOWED_MB=6900; RX_VK_MIRROR_SPARSE=1 against unset (then off) |
the main outdoor area, a scripted five-minute walk | run02 (8 GB allowance): on 59.1-59.3 fps, VRAM peak 6,676-7,018 MB; off 59.0-59.1 fps, 7,028-7,138 MB; mirrors with memory 1,351-1,366 of 1,888 MB |
| 2026-11-12 | Hookshot: v2.1.0-rc.3-249-g564ea648 and later builds of this change (sparse VRAM mirrors, https://github.com/NotRexed/Hookshot/pull/2794) on Hookshot main c544828f (unmerged then); against the last v2.1.0-rc.3 on 6dcc15c7 | Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32; nothing else running, RX_VK_TEST_ALLOWED_MB=6144, RX_VK_MIRROR_MB=1536; RX_VK_MIRROR_SPARSE=1 against unset (then off) |
the main outdoor area, a scripted five-minute walk | run01 (6 GB allowance): three pairs: 57.3 / 55.3 / 58.2 fps against 47.7 / 51.5 / 49.9; frames over 25 ms 625 / 1,160 / 408 against 3,042 / 1,840 / 2,114; VRAM peak 6,394 / 6,200 / 6,387 MB against 7,327 / 6,254 / 6,444. No device lost |
Earlier runs (frozen)
These existing rows keep their original text and order, including their relative references and repeated entries. Do not append new runs here. Frame rates come from the named run’s log at the stated spot, unless the row says otherwise.
| Date | Build | OS / desktop | GPU / driver | CPU / RAM | Settings | Where | Result |
|---|---|---|---|---|---|---|---|
| 2026-11-13 | fix/default-output-2560x1440 b8542e6b (main fbd42024 + this change; longshot v3.0.0-rc.3-16-gb8542e6b): a copy of main’s game folder built again with make_game.py --update, alternated with main |
Windows 11 Pro 26200, a 2560x1440 window in front on a 3840x2160 screen (unattended, input over the control bridge) | RTX 5070 Ti 16 GB, 621.01 | Ryzen 7 7800X3D, 32 GB | No Output resolution chosen (no output_res= line, no HOS_OUTPUT_RES): 2560x1440. Quality preset Custom, Render resolution 75% (1920x1080), game’s own upscaler, Lanczos window filter, frame generation off, no cap (HOS_FPS_MAX=1000), sound off |
Copied saves: the starting forest (child) and the main outdoor area (child), standing, 60 s after 15 s of settling, 3 runs each | oe: output resolution 2560x1440 (none chosen: the game's), nvn: window crop 0,0 2560x1440. Medians of the 3 runs (PresentMon 2.6; GPU time from the log): the starting forest 82.7 fps, 1% low 63.3, frame time 12.04 ms, p99 14.13 ms, GPU 6.9 ms a frame; the main outdoor area 100.8 fps, 1% low 82.3, 9.85 ms, p99 11.54 ms, GPU 6.5 ms. No device lost |
| 2026-11-13 | main fbd42024 (longshot v3.0.0-rc.3-15-gfbd42024), built with make_game, alternated with the row above |
as above | as above | as above | As above; no Output resolution chosen: 2048x1152, so Render resolution 75% is 1536x864 | as above | oe: output resolution 2048x1152 (the game's), nvn: window crop 0,0 2048x1152. the starting forest 81.9 fps, 1% low 64.9, frame time 12.11 ms, p99 14.08 ms, GPU 6.6 ms a frame; the main outdoor area 102.3 fps, 1% low 84.3, 9.76 ms, p99 11.14 ms, GPU 6.1 ms. No device lost |
| 2026-11-14 | fix-split-ub-flush 1c0e220a (https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/189, RX_VK_SPLIT_UBFLUSH default) against its parent on main, fbd42024: both renderer packs (scripts/dev_packs.py) over one base at f647a072, differing only in longshot_vk.dll; the base built without the hot list and profile (the v3.0.0-rc.3 Windows builder has no work folder) |
Windows 11 Pro Insider Preview (26220), 2048x1152 | Intel Arc B390 (Panther Lake iGPU; VRAM mirror of buffer pools off), 32.0.101.9150 |
Core Ultra X7 358H, 32 GB | Default options | the starting forest, the same save, about 2-4 minutes walking between the houses and trees; main, PR, main (test request T-K2T5) | White flashes and flickering textures in both main runs, none with the PR. The over-bright glow on the trees and houses stays with the PR. fps, median of the 5 s windows (first 6 left out): main 21.9 and 21.9, PR 20.5 (one run each way: no speed claim). Main’s runs ended in a GPU hang after 5742 frames (INDIRECT dispatch of 685adb41edc19de7) and a device loss after 2867 (dispatch of 499dbfd0721d2858); the PR’s closed normally after 3991 |
| 2026-11-13 | main 9a186f71 against fix-split-ub-flush 0a5d21e6 (this change on it, https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/189), built from source (llvm-mingw), interleaved with two other Intel PRs’ builds (main, split, pcache, fadd, then again) | Windows 11 Pro (26300), output 1920x1080 | Intel Arc Graphics (Core Ultra 7 155H iGPU, Meteor Lake), 32.0.101.9175 | Core Ultra 7 155H, 32 GB | Render resolution 4, handheld profile, RX_VK_NO_PREDICT=1 on every side, a capture every 60 frames |
A save in the starting forest, then the control bridge’s teleport to the village and Death Mountain Trail, about 130 s each; 3 runs per side | Median over the runs of each run’s median 5 s window (the first 4 after a teleport left out) and GPU time a frame, main against this change: the starting forest 5.5 / 5.4 fps, 160.5 / 168.1 ms (3 runs each); the village 7.4 / 8.1 fps, 120.4 / 107.9 ms (2 / 3); Death Mountain Trail 8.7 / 9.3 fps, 102.0 / 93.6 ms (1 / 1): within the run-to-run spread. Device lost in the village in 2 of 3 runs on both sides. White blocks (bright flat blocks per capture): main 17 captures in 1 run of 3, up to 10.2 % of the picture; this change 0 in 3 runs |
| 2026-11-13 | host and game folder built from main 000f19fb; longshot_vk.so built alone from main fbd42024 and from RX_VK_READBACK_LATE’s change on it, 3a6fbc0c, same flags |
SteamOS 3.9.5 (kernel 7.2.7-valve2), headless | Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel; the GPU’s clock fixed at 1600 MHz (Steam’s performance menu) | AMD Custom APU 0932 (4 cores, 8 threads), 16 GB | Minimum preset (RX_VK_LOW_SPEC=1), handheld profile (HOS_HANDHELD=1), the game’s own upscaler, RX_VK_VRAM_MB=2400, RX_VK_TEX_MAX_DIM=1024, RX_VK_DESC_SLOTGEN=1, RX_VK_LOAD=create, HOS_RESERVE_CORE=0; RX_VK_READBACK_LATE as each arm names it (0 names no rule) |
a save loaded by script, standing still: a heavy view (about 2,150 draws a frame), the main outdoor area (about 1,600), the starting forest (about 2,600) | Five arms alternated in each round, each run measured over exactly 720 presents after 3 log windows: main; the change with no rule; newest; 3980; newest,3980. Medians, fps, in that order: heavy view (3 rounds) 24.28 / 24.30 / 30.13 / 24.86 / 30.33; the main outdoor area (3 rounds) 21.88 / 22.00 / 26.40 / 23.08 / 26.48; the starting forest (2 rounds, their mean) 23.44 / 23.32 / 26.30 / 23.86 / 26.78. A readback that waits takes 2.36 ms on main on the heavy view, 0.01 with newest, 1.58 with 3980, 0.03 with both; in the main outdoor area 4.47 ms with no rule, 0.50 with both. The flash detector with the camera turning (RX_VK_FLASH=1, about 1,800 presents a run, the starting forest and the main outdoor area): 0 flashes with no rule and 0 with newest,3980. No errors in 48 runs, no GPU timeout in the kernel log |
| 2026-11-13 | main d4a69456 + Longshot PR 59 at 8a091b34 (before the reviews’ fixes: the reach 1 unit a frame, poses read under the lock), only that PR | - | RTX 5080, - | - | DLSS Frame Generation 2x, 3x and 4x | The four Gossip Stones by the Temple of Time, the camera moved back and forth (main’s repro) | No ghosting, smearing or copies of the stones seen at any of the three (by eye). Run by Desmond27 |
| 2026-11-13 | main d4a69456 against Longshot PR 59 at 8a091b34 (before the reviews’ fixes), both built natively; interleaved main, PR, main, PR, then the PR with RX_VK_DLSS_ACTOR_LIVE=1 and with RX_VK_DLSS_ACTORS=0 |
Arch Linux (kernel 7.2.15), windowless | RTX 4090 24 GB, 615.78.08 | i9-13900K | 3840x2160, DLSS Transformer (K) Performance, Reflex On + Boost, frame generation off (DLSS-G not available: its Linux library was missing); RX_VK_DLSS_DEBUG=motion |
Child Link, daylight, the four Gossip Stones outside the Temple of Time; scripted left/right passes of 100 game frames, every present captured at 960x540 | Stone motion-vector signature frames: main 0/460 and 0/460, PR 0/460 and 0/460; PR with RX_VK_DLSS_ACTOR_LIVE=1 (the old copy) 2/460. Actor motion CPU a frame (status lines, PC heavily loaded): main 0.097 and 0.094 ms, PR 0.421 and 0.399 ms (LIVE=1 0.112). PR: poses 30714 and 29634 read at the draw, 0 copied later. No device lost. The generated-frame fault itself not checked (FG off). Run by mooki |
| 2026-11-13 | perf/low-vram-mirror-budget 091daa9, a full game build (scripts/dev_packs.py base, llvm-mingw, NVIDIA DLSS SDK); main’s mirrors by RX_VK_MIRROR_MB=512 in it, and 1024 MB the same way; alternated: main, PR, 1024, PR with RX_VK_MIRROR_TIGHT=1, twice |
Windows 11 Home (26200), 2048x1152 | RTX 3060 Laptop 6 GB, 621.14 | Ryzen 7 6800HS, 16 GB | Low (6GB) with Render resolution Native (DLAA) in its Handheld profile: DLSS at 1536x864; frame generation off | the main outdoor area, walking the same route, 75 s a run | fps (two runs each): main 16.9 and 17.7 (950 frames over 50 ms a run), PR 36.2 and 36.1 (about 49), 1024 MB 31.0 and 31.5, PR with the 90% rule 30.5 and 31.4. VRAM used of allowed (median, max): main 80%, 87%; PR 96%, 99%; 1024 MB 89%, 93%; the 90% rule 95%, 99%. No Vulkan errors |
| 2026-11-13 | as above, the renderer built alone from 091daa9 plus an adaptive mirror budget (on backup/adaptive-mirror-budget, not in this PR), run with RX_VK_MIRROR_ADAPT=0 (this PR’s fixed quarter), RX_VK_MIRROR_MB=512 (main’s) and adaptive; alternated, twice |
as above | as above | as above | Docked profile, Render resolution Native (DLAA): DLSS at 2048x1152; and the Low (6GB) preset (1152x648); frame generation off | the main outdoor area, walking the same route, 60 s a run | Docked native: main 12.6 and 12.5 fps (about 780 frames over 50 ms a run, GPU 61 ms), this PR 24.4 and 23.4 (135-153, GPU 31-32 ms), VRAM 97% (max 98%; 5200 of 5063 MB allowed for a moment, nothing failed); the adaptive budget 20.9 and 20.8 (353-406), dropped. Low: this PR 33.2 and 35.2, adaptive 31.9 and 30.8. No Vulkan errors |
| 2026-11-13 | v3.0.0-rc.3 game (source cb01b5a3); its renderer built alone (RX_VK_ONLY, RX_RELEASE, llvm-mingw, NVIDIA DLSS SDK) from main fbd4202 and from perf/low-vram-mirror-budget c17b339, and c17b339 with RX_VK_MIRROR_MB=1024, swapped between runs: main, PR, 1024, main, PR, 1024 |
Windows 11 Home (26200), 2048x1152 | RTX 3060 Laptop 6 GB (Optimus: drives no desktop), 621.14 | Ryzen 7 6800HS, 16 GB (15 GB usable) | Quality preset Low (6GB): its VRAM budget 3.6 GB and its DLSS render size, 1152x648, in every run; frame generation off, no cap | the main outdoor area, walking the same route, about 2 minutes a run | Median of the in-field “presents” lines over both runs: main 17.6 fps (16.7, 18.6), PR 39.2 (39.2, 38.9), 1024 MB 30.7 (22.4, 33.1). From the in-field VRAM: lines (median, max): main 4085 MB used of 4998 allowed (82%, 90%), PR 4818 of 5037 (95%, 98%), 1024 MB 4432 of 5007 (89%, 92%); textures held 1619, 1560 and 1624 MB. Render-target pool “VRAM short: given back”: main 0, PR 5, 1024 MB 0. Mirrors 512, 1498 (a quarter of the 5992 MB the driver reports) and 1024 MB; the largest pool left in system memory 224, 92 and 152 MB. No Vulkan errors |
| 2026-11-13 | fix/evict-writeback-newer-drawing c8e54dc2 (main 514591a3 + this change): the renderer built alone (RX_VK_ONLY, llvm-mingw; v3.0.0-rc.3-26-gc8e54dc2, source 02e9107d) in a game folder of longshot v3.0.0-rc.3-27-gf516654c, alternated with main’s renderer in the same folder |
Windows 11 Pro 26200, a 1707x960 window (unattended, input over the control bridge) | RTX 5070 Ti 16 GB, 621.01 | Ryzen 7 7800X3D, 32 GB | Output resolution 2560x1440 (HOS_OUTPUT_RES), Render resolution 75%, game’s own upscaler, Lanczos window filter, frame generation off, no cap (HOS_FPS_MAX=1000), sound off; RX_VK_TEST_ALLOWED_MB=6144 (a 6 GB card’s allowance) |
A copied save, the debug teleport to the main outdoor area (child), standing 25 s, then the pause menu for 5 s; 3 runs each | The HUD’s labels (d-pad icon, item count, key prompt) drawn right in 3 of 3 runs, no upload over a drawn image line for them; 3, 3 and 2 write-backs skipped, of render targets (1920x1080, 2560x1440, 960x540) unused for 600 frames or more. 104.0, 98.8 and 99.3 fps (medians of the log’s fps lines in the field). One run without the allowance: labels right, 112.1 fps. No device lost |
| 2026-11-13 | main 514591a3: the renderer built alone the same way (v3.0.0-rc.3-25-g514591a3, source 2af40fe2), alternated with the row above |
as above | as above | as above | as above | as above | The labels as noise in 3 of 3 runs, from the area’s load until the pause menu redrew them; 8 upload over a drawn image lines for them in each run, in the frame a 1920x1080 render target over their storage was written back and evicted. 105.5, 95.1 and 100.7 fps. No device lost |
| 2026-11-13 | game folder built from main e51e28e9 (longshot e51e28e9 (source 8dd02aab)); longshot_vk.dll built alone (RX_VK_ONLY, release, llvm-mingw) from fix/virtual-pool-drawn-page-map 4830ea7d (this change on e51e28e9); HUD A/B run 4 of 4 (this PR) |
Windows 11 Pro 26200, fullscreen | RTX 5080 16 GB, 610.88 | Ryzen 9 7900X3D, 64 GB | 2560x1440, DLAA, Reflex off, frame rate cap 90, frame generation off, VRAM budget 11 GB; RX_VK_PARALLEL=4, HOS_PIN=*=FFF; DualSense |
the main outdoor area next to Castle Town by day, walked into Castle Town and its market and back out | The HUD’s labels and D-pad cross drawn at the town gate. No upload over a drawn image line of the bug’s form (dirty, finalized, no “a GPU buffer wrote it”) for the HUD’s textures or render targets; 12 for a 640x360 NVN 0x14 image the texture budget remakes (main has it too). No validation, device-lost or VK_ERROR line; closed from the pause menu |
| 2026-11-13 | game folder built from main e51e28e9 (longshot e51e28e9 (source 8dd02aab)); longshot_vk.dll built alone (RX_VK_ONLY, release, llvm-mingw) from main e51e28e9 (the same file as the full build’s); HUD A/B run 3 of 4 (main) |
Windows 11 Pro 26200, fullscreen | RTX 5080 16 GB, 610.88 | Ryzen 9 7900X3D, 64 GB | 2560x1440, DLAA, Reflex off, frame rate cap 90, frame generation off, VRAM budget 11 GB; RX_VK_PARALLEL=4, HOS_PIN=*=FFF; DualSense |
the main outdoor area next to Castle Town by day, walked into Castle Town and its market and back out | At the town gate the R1 and R2 labels and the D-pad cross are not drawn. upload over a drawn image lines of the bug’s form: the six label and cross textures at one present (5926), 6 other HUD textures, 9 render targets (2560x1440 NVN 0x34 and 0xa among them). No validation, device-lost or VK_ERROR line; closed from the pause menu |
| 2026-11-13 | game folder built from main e51e28e9 (longshot e51e28e9 (source 8dd02aab)); longshot_vk.dll built alone (RX_VK_ONLY, release, llvm-mingw) from fix/virtual-pool-drawn-page-map 4830ea7d (this change on e51e28e9); HUD A/B run 2 of 4 (this PR) |
Windows 11 Pro 26200, fullscreen | RTX 5080 16 GB, 610.88 | Ryzen 9 7900X3D, 64 GB | 2560x1440, DLAA, Reflex off, frame rate cap 90, frame generation off, VRAM budget 11 GB; RX_VK_PARALLEL=4, HOS_PIN=*=FFF; DualSense |
the main outdoor area next to Castle Town by day, walked into Castle Town and its market and back out | The HUD’s labels and D-pad cross drawn at the town gate. No upload over a drawn image line of the bug’s form for the HUD’s textures or render targets; 11 for the 640x360 NVN 0x14 image the texture budget remakes. No validation, device-lost or VK_ERROR line; closed from the pause menu |
| 2026-11-13 | game folder built from main e51e28e9 (longshot e51e28e9 (source 8dd02aab)); longshot_vk.dll built alone (RX_VK_ONLY, release, llvm-mingw) from main e51e28e9 (the same file as the full build’s); HUD A/B run 1 of 4 (main) |
Windows 11 Pro 26200, fullscreen | RTX 5080 16 GB, 610.88 | Ryzen 9 7900X3D, 64 GB | 2560x1440, DLAA, Reflex off, frame rate cap 90, frame generation off, VRAM budget 11 GB; RX_VK_PARALLEL=4, HOS_PIN=*=FFF; DualSense |
the main outdoor area next to Castle Town by day, walked into Castle Town and its market and back out | At the town gate the R1 and R2 labels and the D-pad cross are not drawn, on the first try. upload over a drawn image lines of the bug’s form: the six label and cross textures at one present (7002), 4 other HUD textures, 4 render targets. No validation, device-lost or VK_ERROR line; closed from the pause menu |
| 2026-11-13 | game v3.0.0-rc.3-4-gfdc2840b (main fdc2840b, --fsr); renderer built alone (RX_VK_ONLY, llvm-mingw, LONGSHOT_FSR=1) from fix/psychov35-fsr4 ea988050 |
Windows 11 Pro 26300, fullscreen | RX 9070 XT 16 GB, Adrenalin 26.10.2 | Ryzen 7 9800X3D, 32 GB | 2560x1440, HDR10 (white 200, peak 700 nits), HDR Native, HDR tone mapper Vanilla, AMD FSR 4 Native AA (FSR 4.1.2), Sharpness 50; RX_VK_PSYCHOV35_DLSS=1 |
Forest Temple, indoors; cold start, about a minute standing | 600 of 600 FSR 4 frames in 6 of 6 complete intervals; no PsychoV35 line (default path unchanged) |
| 2026-11-13 | game v3.0.0-rc.3-4-gfdc2840b (main fdc2840b, --fsr); renderer built alone (RX_VK_ONLY, llvm-mingw, LONGSHOT_FSR=1) from fix/psychov35-fsr4 ea988050 |
Windows 11 Pro 26300, fullscreen | RX 9070 XT 16 GB, Adrenalin 26.10.2 | Ryzen 7 9800X3D, 32 GB | 2560x1440, HDR10 (white 200, peak 700 nits), HDR Native, HDR tone mapper PsychoV35, AMD FSR 4 Native AA (FSR 4.1.2), Sharpness 50; RX_VK_PSYCHOV35_DLSS=1 |
Forest Temple, indoors; cold start, about a minute standing | PsychoV35 FSR 4: RGBA16F linear route, FSR 4 made for HDR, linear colour; 600 of 600 FSR 4 frames in 8 of 8 complete intervals. Smooth edges, Sharpness row works, HDR brightness right (#226 fixed) |
| 2026-11-13 | game v3.0.0-rc.3-4-gfdc2840b (main fdc2840b, --fsr); renderer built alone (RX_VK_ONLY, llvm-mingw, LONGSHOT_FSR=1) from fix/psychov35-fsr4 ea988050 |
Windows 11 Pro 26300, fullscreen | RX 9070 XT 16 GB, Adrenalin 26.10.2 | Ryzen 7 9800X3D, 32 GB | 2560x1440, HDR10 (white 200, peak 700 nits), HDR Native, HDR tone mapper PsychoV35, AMD FSR 4 Native AA (FSR 4.1.2), Sharpness 50; RX_VK_PSYCHOV35_DLSS=0 |
Forest Temple, indoors; cold start, about a minute standing | the game's upscale runs (no scene camera this frame); 0 FSR 4 frames in every interval. Jagged, oversharpened, Sharpness row without effect (as on main) |
| 2026-11-13 | game v3.0.0-rc.3-4-gfdc2840b (main fdc2840b, --fsr); renderer built alone (RX_VK_ONLY, llvm-mingw, LONGSHOT_FSR=1) from fix/psychov35-fsr4 1ba927d0 (the same change on #108’s branch) |
Windows 11 Pro 26300, fullscreen | RX 9070 XT 16 GB, Adrenalin 26.10.2 | Ryzen 7 9800X3D, 32 GB | 2560x1440, HDR10 (white 200, peak 700 nits), HDR Native, HDR tone mapper PsychoV35, AMD FSR 4 Native AA (FSR 4.1.2), Sharpness 100; RX_VK_PSYCHOV35_DLSS=0 |
Forest Temple, indoors; cold start, about a minute standing | As the run above: the game’s upscale, 0 FSR 4 frames |
| 2026-11-13 | game v3.0.0-rc.3-4-gfdc2840b (main fdc2840b, --fsr); renderer built alone (RX_VK_ONLY, llvm-mingw, LONGSHOT_FSR=1) from fix/psychov35-fsr4 1ba927d0 (the same change on #108’s branch) |
Windows 11 Pro 26300, fullscreen | RX 9070 XT 16 GB, Adrenalin 26.10.2 | Ryzen 7 9800X3D, 32 GB | 2560x1440, HDR10 (white 200, peak 700 nits), HDR Native, HDR tone mapper PsychoV35, switched to Vanilla and back during play, AMD FSR 4 Native AA (FSR 4.1.2), Sharpness 100; RX_VK_PSYCHOV35_DLSS=1; HDR Native / Extended / Native during play |
Forest Temple, indoors; cold start, about a minute standing, then moving between areas | FSR 4 frames: 19 of 27 complete intervals at 600; 595 to 599 around the tone mapper and HDR switches (1 to 3 frames of unexpected scene colour format each, then the route again), 532 and 522 after an area change, one 351 not explained by the log; one interval of FSR 4 (not Native AA) at 600 |
| 2026-11-13 | host and game folder built from main 000f19fb; longshot_vk.so built alone from main 000f19fb (before) and from this change on it, ffee45f4 (after), same flags |
SteamOS 3.9.5 (kernel 7.2.7-valve2), headless | Steam Deck OLED: AMD Custom GPU 0932 (RADV VANGOGH), Mesa 26.2.0-devel | AMD Custom APU 0932 (4 cores, 8 threads), 16 GB | Minimum preset (RX_VK_LOW_SPEC=1), handheld profile (HOS_HANDHELD=1), the game’s own upscaler, RX_VK_VRAM_MB=2400, RX_VK_TEX_MAX_DIM=1024, RX_VK_DESC_SLOTGEN=1, RX_VK_LOAD=create, HOS_RESERVE_CORE=0 |
a save loaded by script, standing still on a heavy view (about 2,500 draws a frame); the main outdoor area (about 1,600 draws) | Renderers alternated, each run measured over exactly 720 presents after 3 log windows; the render thread’s cycles and instructions by perf stat on that thread (user space). Heavy view, 4 runs each, medians: 77.04 → 73.57 M instructions and 88.30 → 86.08 M cycles a frame, replay 31.67 → 30.45 ms, 23.83 → 23.35 fps (inside the runs’ spread, 23.3 to 24.3 on both sides: the frame is about 42 ms, the rest is waits). the main outdoor area, 2 runs each: 47.01 → 44.70 M instructions, 60.65 → 58.85 M cycles, replay 19.95 → 19.43 ms, 22.58 → 22.61 fps (the GPU’s 24.7 ms a frame is the limit). No errors in 12 runs |
| 2026-11-13 | v3.0.0-rc.3 game (source cb01b5a3); its renderer built alone (RX_VK_ONLY, RX_RELEASE, llvm-mingw, NVIDIA DLSS SDK) from main 0edc137 and from perf/low-vram-mirror-budget d27a462, swapped between runs: main, PR, main, PR |
Windows 11 Home (26200), 2048x1152 | RTX 3060 Laptop 6 GB, 621.14 | Ryzen 7 6800HS, 16 GB (15 GB usable) | Quality preset Auto (this PC), VRAM budget 3.6 GB (Low), NVIDIA DLSS at 33% (683x384) in every run, frame generation off, no cap; no RX_VK_MIRROR_MB |
the main outdoor area, walking, about 2 minutes a run | Median of the in-field “presents” lines: main 19.5 and 21.4 fps (20.1 over both), mirrors 512 MB, a 224 MB pool “stays in system memory”; PR 43.9 and 41.4 fps (42.2 over both), mirrors 1498 MB (a quarter of the 5992 MB the driver reports, which also caps Low’s 1536), the largest pool left out 24 MB. No Vulkan errors in either |
| 2026-11-13 | v3.0.0-rc.3 (source cb01b5a3), the launcher’s build, with RX_VK_MIRROR_MB=512 (main’s mirrors for a 6 GB card on the Low (3.6 GB) VRAM budget) against RX_VK_MIRROR_MB=1536 (perf/low-vram-mirror-budget’s), alternated: main, PR, main, PR |
Windows 11 Home (26200), 2048x1152 | RTX 3060 Laptop 6 GB, 621.14 | Ryzen 7 6800HS, 16 GB (15 GB usable) | Quality preset Auto (this PC), VRAM budget 3.6 GB, NVIDIA DLSS, Render resolution settled at 33% (683x384) in every run (the first main run started at 50% and stepped down), frame generation off, no cap | the main outdoor area, walking, about 2 minutes a run | Median of the in-field “presents” lines: main 17.7 and 20.7 fps (18.8 over both), GPU frame 34-37 ms, a 224 MB pool “stays in system memory” (mirrors 444-488 MB of 512); PR 36.3 and 39.5 fps (38.6 over both), GPU frame 14 ms, 114-120 pools mirrored (1536 MB) |
| 2026-11-13 | Renderer 6b8bc562 / cdee3127; host 7afe6ae1 / 7cfff061; pr-psycho-dlaa | Windows 11 Home 26300, windowed | RTX 4070 12 GB, 621.01 | i7-13700F, 32 GB | 2560x1440 output, 2560x1421 window; DLAA 2560x1440, K, sharpness 0; Native HDR, PsychoV35; white 200 / peak 400 nits for HDR; NVIDIA FG 2x; RX_VK_PSYCHOV35_DLSS=1 | Copied the main outdoor area child save, cold start, 40 s standing after settling | Last two complete 600-frame intervals: 600/600 DLSS/DLAA evaluations and NVIDIA FG sets; zero missing-input fallbacks. Correctness counts, no FPS claim |
| 2026-11-13 | Renderer 6b8bc562 / cdee3127; host 7afe6ae1 / 7cfff061; pr-psycho-on2 | Windows 11 Home 26300, windowed | RTX 4070 12 GB, 621.01 | i7-13700F, 32 GB | 2560x1440 output, 2560x1421 window; DLSS Quality 1707x960, K, sharpness 0; Native HDR, PsychoV35; white 200 / peak 400 nits for HDR; NVIDIA FG 2x; RX_VK_PSYCHOV35_DLSS=1 | Copied the main outdoor area child save, cold start, 40 s standing after settling | Last two complete 600-frame intervals: 600/600 DLSS/DLAA evaluations and NVIDIA FG sets; zero missing-input fallbacks. Correctness counts, no FPS claim |
| 2026-11-13 | Renderer dd02c6cc / cb01b5a3; host 7afe6ae1 / 7cfff061; main-psycho-off2 | Windows 11 Home 26300, windowed | RTX 4070 12 GB, 621.01 | i7-13700F, 32 GB | 2560x1440 output, 2560x1421 window; DLSS Quality 1707x960, K, sharpness 0; Native HDR, PsychoV35; white 200 / peak 400 nits for HDR; NVIDIA FG 2x; switch unset | Copied the main outdoor area child save, cold start, 40 s standing after settling | Last two complete 600-frame intervals: 0/600 NVIDIA FG, 600 missing-input fallbacks; game network used. Correctness counts, no FPS claim |
| 2026-11-13 | Renderer 6b8bc562 / cdee3127; host 7afe6ae1 / 7cfff061; pr-vanilla-sdr | Windows 11 Home 26300, windowed | RTX 4070 12 GB, 621.01 | i7-13700F, 32 GB | 2560x1440 output, 2560x1421 window; DLSS Quality 1707x960, K, sharpness 0; SDR, Vanilla; white 200 / peak 400 nits for HDR; NVIDIA FG 2x; switch unset | Copied the main outdoor area child save, cold start, 40 s standing after settling | Last two complete 600-frame intervals: 600/600 DLSS/DLAA evaluations and NVIDIA FG sets; zero missing-input fallbacks. Correctness counts, no FPS claim |
| 2026-11-13 | Renderer dd02c6cc / cb01b5a3; host 7afe6ae1 / 7cfff061; main-vanilla-sdr | Windows 11 Home 26300, windowed | RTX 4070 12 GB, 621.01 | i7-13700F, 32 GB | 2560x1440 output, 2560x1421 window; DLSS Quality 1707x960, K, sharpness 0; SDR, Vanilla; white 200 / peak 400 nits for HDR; NVIDIA FG 2x; switch unset | Copied the main outdoor area child save, cold start, 40 s standing after settling | Last two complete 600-frame intervals: 600/600 DLSS/DLAA evaluations and NVIDIA FG sets; zero missing-input fallbacks. Correctness counts, no FPS claim |
| 2026-11-13 | Renderer dd02c6cc / cb01b5a3; host 7afe6ae1 / 7cfff061; main-psycho-off | Windows 11 Home 26300, windowed | RTX 4070 12 GB, 621.01 | i7-13700F, 32 GB | 2560x1440 output, 2560x1421 window; DLSS Quality 1707x960, K, sharpness 0; Native HDR, PsychoV35; white 200 / peak 400 nits for HDR; NVIDIA FG 2x; switch unset | Copied the main outdoor area child save, cold start, 40 s standing after settling | Last two complete 600-frame intervals: 0/600 NVIDIA FG, 600 missing-input fallbacks; game network used. Correctness counts, no FPS claim |
| 2026-11-13 | Renderer 6b8bc562 / cdee3127; host 7afe6ae1 / 7cfff061; pr-psycho-on | Windows 11 Home 26300, windowed | RTX 4070 12 GB, 621.01 | i7-13700F, 32 GB | 2560x1440 output, 2560x1421 window; DLSS Quality 1707x960, K, sharpness 0; Native HDR, PsychoV35; white 200 / peak 400 nits for HDR; NVIDIA FG 2x; RX_VK_PSYCHOV35_DLSS=1 | Copied the main outdoor area child save, cold start, 40 s standing after settling | Last two complete 600-frame intervals: 600/600 DLSS/DLAA evaluations and NVIDIA FG sets; zero missing-input fallbacks. Correctness counts, no FPS claim |
| 2026-11-13 | Renderer 6b8bc562 / cdee3127; host 7afe6ae1 / 7cfff061; pr-vanilla-hdr | Windows 11 Home 26300, windowed | RTX 4070 12 GB, 621.01 | i7-13700F, 32 GB | 2560x1440 output, 2560x1421 window; DLSS Quality 1707x960, K, sharpness 0; Native HDR, Vanilla; white 200 / peak 400 nits for HDR; NVIDIA FG 2x; switch unset | Copied the main outdoor area child save, cold start, 40 s standing after settling | Last two complete 600-frame intervals: 600/600 DLSS/DLAA evaluations and NVIDIA FG sets; zero missing-input fallbacks. Correctness counts, no FPS claim |
| 2026-11-13 | Renderer dd02c6cc / cb01b5a3; host 7afe6ae1 / 7cfff061; main-vanilla-hdr | Windows 11 Home 26300, windowed | RTX 4070 12 GB, 621.01 | i7-13700F, 32 GB | 2560x1440 output, 2560x1421 window; DLSS Quality 1707x960, K, sharpness 0; Native HDR, Vanilla; white 200 / peak 400 nits for HDR; NVIDIA FG 2x; switch unset | Copied the main outdoor area child save, cold start, 40 s standing after settling | Last two complete 600-frame intervals: 600/600 DLSS/DLAA evaluations and NVIDIA FG sets; zero missing-input fallbacks. Correctness counts, no FPS claim |
| 2026-11-13 | fix/occlusion-queries 3f67aad0 (main 839e85ff + this change: the renderer built alone and put in a copy of main’s folder) against main 839e85ff, alternated | Linux (CachyOS, kernel 7.3.3), a private headless sway 2048x1152, scripted input, no sound device | RTX 3090 24 GB, 620.70.06 | Ryzen 9 7900X, 30 GB (swapping in every run) | Default options, shaders prepared, cap 240; RX_VK_OCC_LOG=1 both sides |
The pristine starting-area save: the starting forest standing 40 s, then the camera turning at full speed 30 s; Castle Town Market (debug teleport) the same; 4 runs each | Medians of the runs’ medians, main / this change. starting-area standing: 2,862 / 2,632 draws, GPU 10.1 / 10.1 ms, frame 16.54 / 15.95 ms. starting-area turning: 1,928 / 1,767 draws, GPU 9.00 / 9.25 ms, frame 14.43 / 14.35 ms. Market standing: 2,926 / 2,464 draws, GPU 8.60 / 8.07 ms, frame 16.55 / 15.34 ms. Market turning: GPU 5.38 / 5.42 ms, frame 12.84 / 12.82 ms, longest frame 71 / 46 ms. 22-60% of reports read 0, written 1-4 recordings after the report, 0 left visible by a draw that did not run. Nothing missing in screenshots at the same spot |
| 2026-11-13 | fix/occlusion-visible-unless-known (main fdc2840b + this change: the renderer built alone, RX_VK_ONLY, as main’s for the comparison, in copies of one main folder) against main fdc2840b, alternated |
Linux (CachyOS, kernel 7.3.3), a private headless sway 2048x1152, scripted input, no sound device | RTX 3090 24 GB, 620.70.06 | Ryzen 9 7900X, 30 GB (swapping in every run) | Default options, shaders prepared, cap 240; RX_VK_OCC_LOG=1 both sides |
The pristine starting-area save: the starting forest standing 40 s, then turning 30 s; Castle Town Market (debug teleport) the same; 4 runs each (P M M P, then M X M X with X the same code), 2 with RX_VK_OCC_QUERIES=0 |
Medians of the runs’ values, main / this change / queries off. starting-area standing: 2,600 / 2,614 / 2,864 draws, frame 15.47 / 15.20 / 15.32 ms. Market standing: 2,433 / 2,440 / 2,928 draws, GPU 7.53 / 7.40 / 7.95 ms, frame 15.67 / 16.12 / 17.34 ms. Market turning: frame 13.36 / 13.60 / 13.05 ms. Validation layer, one short session each: main 02741 and 00807 to the cap, this change none. 0 flashes either side. 0 queries left unbegun in play |
| 2026-11-12 | fix/load-wait-rule 3a297868 (main 4dd972a8 + this change; the commit before the merge of main that brought the rename: same code, old hookshot_ file names), with a test-only switch added that is not part of the change. The merged branch has not run in a game |
Linux, a private headless sway, scripted input | RTX 3090, - | - | A folder of its own prepared by --prepare-shaders, RX_VK_LOAD=create; nothing injected |
A short session: the first area, a teleport, a second area; three sessions (the first start after a preparation, one between the injected ones below, the first start after the second preparation) | Waits the rule counts (over 20 ms): 27 ms (8 first uses met a load, 31 ms in all), 0 ms (23 first uses, 32 ms in all, the longest 5.9 ms), 0 ms (11 first uses, 6.6 ms in all). No line from the safety net; no marker (longshot_load_start.txt) and no .waited file after any of them |
| 2026-11-12 | as above | as above | as above | as above | as above, with long waits injected: the test-only switch, which is not part of the change, makes the first N first uses that meet an unfinished load take a set time longer, inside the time the safety net measures. The rule’s own code ran unchanged; no long wait could be produced on this PC without it | as above, six sessions in the order given; “next start” is a start without RX_VK_LOAD, to see what the default picks |
One wait of 300 ms: no line, no file, next start create. Two of 300 ms (601 ms counted): “over 500 ms. Once is let pass”, .waited left. One of 300 ms right after: no line, no file (the strike was taken and not renewed), next start create. Two of 300 ms (600 ms): “Once is let pass”, .waited. Waits of 600 ms right after: “over 500 ms, as the session before this one was”, the marker, next start start. Two of 300 ms with the marker there (600 ms): “Once is let pass”, the marker and .waited |
| 2026-11-12 | as above | as above | as above | as above | The marker and .waited both in the folder |
A preparation (--prepare-shaders), no game |
Both files removed; the next start picks create |
| 2026-11-12 | as above | as above | as above | as above | RX_VK_LOAD=create, one wait of 2,100 ms injected by the same test-only switch (not part of the change) |
A short session, as above | 2,136 ms counted: “over 2000 ms, so the next start loads every listed program before the title again”, the marker; the next start picks start |
| 2026-11-12 | as above | as above | as above | as above | RX_VK_LOAD=create, no load workers (RX_VK_LOAD_WORKERS=0), nothing injected |
A short session, as above | 0 ms counted: 1,157 first uses ran their load themselves, 1,743 ms in all, the longest 16.2 ms. No line, no file (the rule before this change, over 500 ms in all, would have left the marker) |
| 2026-11-12 | main 4dd972a8 + this change, HOS_VIDEO_INPUT_LIMIT=0 (as main) against the default (16) |
Linux (CachyOS, kernel 7.3.3), a private headless sway, scripted input, no sound device (HOS_AUDIO=0) |
RTX 3090 24 GB, 620.70.06 | Ryzen 9 7900X, 30 GB (the PC was swapping in every run) | 2048x1152, default options, cap 60 | A new game from an empty data folder: the first movie (A after 11 s, about 820 pictures of it), then the second (9,254 pictures) to its end | Three runs each, interleaved. Pictures skipped, first movie: 244, 247, 251 without the limit; 8, 198, 1 with it. Second movie: 123, 700, 98 without; 0, 502, 0 with. In the second run of each side the PC swapped hard and frames were late (177 and 146 frames over 33 ms; 0 to 2 in the others): those skips are the late frames’. Waits of over 90 ms between two pictures in a movie’s first 5 s: 44, 43, 43 and 20, 21, 18 without; 1, 0, 0 and 0, 9, 0 with. The second movie ends within 12 ms of its clock on both sides |
| 2026-11-13 | Hookshot fix/commit-limit 905cb5e1 (NotRexed/Hookshot#2972, ported here as fix/commit-limit), against Hookshot main e0e60920, alternated | Windows | RTX 4060 Ti, - | -, 31 GB (115 GB commit limit) | 3440x1440, muted, the commit rule idle (default RX_VK_COMMIT_RESERVE_MB) |
Route run: starting-area, the main outdoor area, the village, 2 runs each | This change: 53 and 57 fps median, p99 40.5 and 36.1 ms, private commit at end 28.9 and 28.6 GB, no commit limit near line (the rule never fired). Main: 51 and 50 fps, p99 43.2 and 42.1 ms, 33.4 and 34.8 GB. Run-to-run variation, not the change: it costs nothing when idle |
| 2026-11-13 | as above, this change only | as above | as above | as above | as above, RX_VK_COMMIT_RESERVE_MB=38397 (the rule forced, about 25 GB of the game’s commit before it kicks in) |
as above, 1 run | 54 fps median, p99 38.7 ms, max 881 ms (the run before, rule not firing: 58, 35.6, 881). 52 commit limit near lines, texture budget down by the shortfall each time (e.g. 3661 to 2812 MB), headroom 37.4 to 38.4 GB; private commit peak 26.7 GB (30.3 without the rule), VRAM blocks peak 9.0 GB (12.9). The 5 s wait at the limit itself not reached |
| 2026-11-13 | v2.1.0-rc.3-313-g9b6072c + this PR, the split-submission fix (#189) and compute stages let go of the 32-lane pin as https://github.com/NotRexed/Hookshot/pull/2564 does, all three in each of the 12 runs (one build), for https://github.com/NotRexed/Hookshot/issues/999; the same patch rebased on main e0e6092 for the PR, built from source (llvm-mingw) | Windows 11 Pro (26300), 1920x1080, scripted input | Intel Arc Graphics (Core Ultra 7 155H iGPU, Meteor Lake), 32.0.101.9175 | Core Ultra 7 155H, 32 GB | Render resolution 4, handheld profile; RX_VK_PASS_TIMING=2, a screenshot every 60 frames; old/new alternated, the game closed normally |
the starting forest outside the player character’s house, standing, 300 s | No igvk64.dll+0x79ddf3 crash in 12 runs (main 9b6072c without them: 4 of 6): the three changes together, not this one alone. The background save ran 3 times and was put off once. Device lost in compute programs in 4 of 12, as on plain main |
| 2026-11-13 | v2.1.0-rc.3-313-g9b6072c + this PR, the pipeline cache fetch lock (https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/190) and compute stages let go of the 32-lane pin as https://github.com/NotRexed/Hookshot/pull/2564 does, one build for every run, for https://github.com/NotRexed/Hookshot/issues/999; the same patch rebased on main e0e6092 for the PR, built from source (llvm-mingw) | Windows 11 Pro (26300), 1920x1080, scripted input | Intel Arc Graphics (Core Ultra 7 155H iGPU, Meteor Lake), 32.0.101.9175 | Core Ultra 7 155H, 32 GB | Render resolution 4, handheld profile; RX_VK_PASS_TIMING=2, a screenshot every 60 frames; old/new alternated, the game closed normally; old: the same build with RX_VK_SPLIT_UBFLUSH=1 |
the starting forest outside the player character’s house, standing, 300 s | White blocks over the forest in 9, 7 and 11 frames of the old runs that ran on, 0 in every new run (an earlier series of this change alone: 7, 1, 1 against 0, 0, 0); fps 4.7-5.7 old, 5.7-6.0 new. Device lost in compute programs (ab7699ee, aba5e6af, 499dbfd0) in 2 of 8 old and 5 of 8 new runs over both series. The same programs lose it with the float-add loop on main and the other Intel builds (12 of 18 runs in https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/191’s series; 499dbfd0 is a float add to a storage buffer, ab7699ee likely one), none with https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/191’s native add (0 of 6), and none in 6 route runs of a build with this change and https://git.magicbean.dev/MagicBeanDev/Longshot/pulls/191 together (13/11): most likely the loop’s, not this change’s, though 5 of 8 against 2 of 8 is not explained |
| 2026-11-13 | v2.1.0-rc.3-330-g82de567a game (main); its renderer built alone (RX_VK_ONLY, llvm-mingw, NVIDIA DLSS SDK) from main d4a69456 and from fix/dlss-preset-settle e8c02084, swapped between runs |
Windows 11 Pro (26300), headless (offscreen 3440x1440) | RTX 5080 16 GB, 621.24 | - | Upscaler NVIDIA DLSS, Render resolution Native (DLAA), output 3440x1440, frame rate cap 240, frame generation off; the DLSS model row set through the control bridge: K, then M and L 11 frames apart, then M and K, then M alone | the starting forest (child, the debug teleport’s arrival), standing; 4 runs, before/after/after/before | Steady, both builds: K 7.4-7.6 ms of GPU a frame (80-84 fps, 140-145 W), L 10.6-10.7 ms (66-67 fps, 198-202 W), M 9.7-9.9 ms (70-72 fps, 186-188 W); the last before run’s scene lighter (K 6.8, L 9.6 ms). K to M to L: before, features for M and L, DLSS’s own 1419 MB, VRAM up 1254 and 1143 MB in 15 s (nvidia-smi); after, L’s alone, 813 MB, up 637 and 637 MB. L to M to K: before up 806 and 827 MB, after 126 and 189 MB. M alone: 657-722 MB both. Issue #17 |
| 2026-11-13 | main @ 316f9319; historical hold-to-toggle run 1 |
NixOS 26.11 / Linux 7.2.7; private headless Sway 1.12 + Xwayland | RX 7900 XT, RADV Mesa 26.2.3 | Ryzen 5 7600X, 32 GB | 1920×1080, 100% render resolution, FG/HDR Off, Uncapped, VSync Auto; stock; mesh fallback, udmabuf, audio off | Castle Town wooden door, copied save | Last five presents readings: 54.1, 55.4, 54.1, 53.6, 53.7; median 54.10 FPS. Same copied save, separate warm-up. Recorded before the review fixes; not a measurement of the current head. |
| 2026-11-13 | this PR @ 0f4b3374; historical hold-to-toggle run 2 |
NixOS 26.11 / Linux 7.2.7; private headless Sway 1.12 + Xwayland | RX 7900 XT, RADV Mesa 26.2.3 | Ryzen 5 7600X, 32 GB | 1920×1080, 100% render resolution, FG/HDR Off, Uncapped, VSync Auto; ZL Toggle On; no held button during FPS samples; mesh fallback, udmabuf, audio off | Castle Town wooden door, copied save | Last five presents readings: 54.4, 53.7, 53.1, 53.0, 52.8; median 53.10 FPS. Same copied save, separate warm-up. Recorded before the review fixes; not a measurement of the current head. |
| 2026-11-13 | main @ 316f9319; historical hold-to-toggle run 3 |
NixOS 26.11 / Linux 7.2.7; private headless Sway 1.12 + Xwayland | RX 7900 XT, RADV Mesa 26.2.3 | Ryzen 5 7600X, 32 GB | 1920×1080, 100% render resolution, FG/HDR Off, Uncapped, VSync Auto; stock; mesh fallback, udmabuf, audio off | Castle Town wooden door, copied save | Last five presents readings: 54.7, 53.9, 53.7, 54.5, 53.9; median 53.90 FPS. Same copied save, separate warm-up. Recorded before the review fixes; not a measurement of the current head. |
| 2026-11-13 | this PR @ 0f4b3374; historical hold-to-toggle run 4 |
NixOS 26.11 / Linux 7.2.7; private headless Sway 1.12 + Xwayland | RX 7900 XT, RADV Mesa 26.2.3 | Ryzen 5 7600X, 32 GB | 1920×1080, 100% render resolution, FG/HDR Off, Uncapped, VSync Auto; ZL Toggle On; no held button during FPS samples; mesh fallback, udmabuf, audio off | Castle Town wooden door, copied save | Last five presents readings: 54.4, 54.2, 54.1, 53.8, 54.7; median 54.20 FPS. Same copied save, separate warm-up. Recorded before the review fixes; not a measurement of the current head. |
| 2026-11-13 | rev137-bc0c1292 (source f94c58fd); Git bc0c1292, hold-to-toggle with main f3fe1763 merged |
NixOS 26.11, Linux 7.2.7; private headless Sway 1.12 + Xwayland | RX 7900 XT, RADV Mesa 26.2.3 | Ryzen 5 7600X, 32 GB | 1920×1080, render 100%, game upscaler, FG/HDR Off, Native mouse, Toggle ZR/ZL On, mesh fallback, udmabuf, audio off; copied save | the castle Town (Market_Child), indoors by the wooden door |
10m06s functional session, clean control-socket exit. ZR press at frames 21466–21481 stayed latched after release; 20 relative XTest mouse events of +20 horizontal counts moved the slingshot aim from the wall to the door without firing (14 seeds before/after; queue empty at frame 23052). ZL press at 18803–18818 held the shield with queue empty at 18946; press at 18954–18969 released it. Opening Insert cleared a later shield latch. Inspected Controls: ZL/ZR On, Plus/Minus disabled, revised help. No crash/device-lost/hang marker. No FPS comparison: another isolated CPU build ran concurrently. |
| 2026-11-13 | feat/mono-audio review follow-up on main d9648863; mono-output-review (source 528a52f6) |
NixOS 26.11 / Linux 7.2.7, private headless Sway/Xwayland | RX 7900 XT 20 GB, RADV Mesa 26.2.3 | Ryzen 5 7600X, 32 GB | 1920×1080, 100% render resolution, game upscaler, FG/HDR Off, Uncapped, VSync Auto; compute mesh fallback; HOS_AUDIO=0; shader prediction budget 60 s |
Castle Town wooden door, copied save, stationary; one live Stereo → Mono → Stereo → Mono run, 20 s per sample after settling | Mono: 0 unequal L/R frames out of 962,048 / 963,072, with 959,216 / 960,560 non-silent frames. Stereo: 960,175 / 960,120 unequal out of 962,048 each. Last-five native FPS medians: Stereo 56.0 / 56.6, Mono 56.6 / 56.4; medians of the two 56.3 / 56.5 (one scene, not a general cost claim). Audio output row on Mono and its live/restart help inspected in a screenshot; saved audio_output=3; no separate Mono row. |
| 2026-11-13 | rev138-2f42b40e01e3 (source a5035217), fix/damage-taken-review 2f42b40e |
NixOS 26.11, Linux 7.2.7; private headless Sway 1.12/Xwayland | RX 7900 XT 20 GB, RADV Mesa 26.2.3 | Ryzen 5 7600X, 32 GB | 1920x1080, game’s upscaler, FG/HDR Off; mesh fallback, audio disabled; copied save | Castle Town roof; first gate-harness attempt | Excluded from gate proof: a three-quarter hit removed 12 (16 to 4) with Quadruple saved, but the harness did not record Debug-tools state at impact. The later run below records it at the damage hook. No product failure claimed from this timing attempt |
| 2026-11-13 | rev138-2f42b40e01e3 (source a5035217), fix/damage-taken-review 2f42b40e |
NixOS 26.11, Linux 7.2.7; private headless Sway 1.12/Xwayland | RX 7900 XT 20 GB, RADV Mesa 26.2.3 | Ryzen 5 7600X, 32 GB | 1920x1080, game’s upscaler, FG/HDR Off; mesh fallback, audio disabled; copied save | Castle Town roof; paired controlled falls | With Quadruple saved, tools Off: raw hit 3, life 16 to 13; tools On again: raw hit 3, life 13 to 1. Quarter-heart units. Read-only hardware breakpoint confirmed tools state and saved choice at each hit; bridge read life after impact. Two setup attempts that did not rise were excluded. Screenshots inspected; no FPS claim |
| 2026-11-13 | rev138-2f42b40e01e3 (source a5035217), fix/damage-taken-review 2f42b40e |
NixOS 26.11, Linux 7.2.7; private headless Sway 1.12/Xwayland | RX 7900 XT 20 GB, RADV Mesa 26.2.3 | Ryzen 5 7600X, 32 GB | 1920x1080, game’s upscaler, FG/HDR Off; Debug tools On, Heart protection Off for measured damage; mesh fallback, audio disabled; copied save | Castle Town roof; four controlled four-quarter hits | Normal: 16 to 12; Double: 12 to 4; Half: 4 to 2, then 2 to 0 and the death animation. Raw hit 4 confirmed at every impact by read-only hardware breakpoint; no guest-memory writes. Two protected height-calibration falls (raw 5 and 4) left life at 16. Screenshots inspected; no FPS claim |
| 2026-11-12 | v2.1.0-rc.3-281-g06ad86fe (source 8d4d275a): main ba3d3f73 + the change of Longshot PR 59 (its first version, before the reviews’ fixes), NotRexed/Hookshot PR 2787’s layouts and the close-up depth test; against v2.1.0-rc.3-271-gba3d3f73, main | Windows 11 Pro, 3840x2160 output | RTX 5090 32 GB, 621.24 | Ryzen 7 9800X3D | DLAA, frame generation 3x by DLSS (NGX), Reflex On + Boost, V-Sync Auto (FIFO) | Castle Town, the four Gossip Stones by the Temple of Time: running past them while the camera turns, about 30 s | Main: the stones break apart in made frames (a stone drawn part way to its neighbour); with RX_VK_DLSS_ACTORS=0 they stay whole. With the change: whole. Every pose read at the draw (“poses N read at the draw, 0 copied later”), 51-74 skinned draws a frame, actor motion 0.15-0.28 ms of CPU a frame, DLSS-G made 590 of 600 presents in play. No device lost |
| 2026-11-12 | d4a69456 + PsychoV35 DLSS port, renderer source 878c957c; host 7afe6ae1; DLAA cold start | Windows 11 Home, build 26300 | RTX 4070 12 GB, NVIDIA 621.01 | i7-13700F, 32 GB | 2560x1440 DLAA (RX_VK_DLSS=2), K; Native HDR PsychoV35, white 200 / peak 400 nits; sharpness 0; NVIDIA FG 2x; RX_VK_PSYCHOV35_DLSS=1 | Copied the main outdoor area child save, 30 s standing after settling | RGBA16F input/output, HDR feature; repeated 600/600 DLAA evaluations and 600/600 NVIDIA FG sets with zero missing-input fallbacks in steady intervals. Method and limits |
| 2026-11-12 | Same renderer and host; extra tone-mapper transition run | Windows 11 Home, build 26300 | RTX 4070 12 GB, NVIDIA 621.01 | i7-13700F, 32 GB | 2560x1440 output, DLSS Quality K; Native HDR PsychoV35, white 200 / peak 400 nits; sharpness 0; NVIDIA FG 2x; switch on | Same copied save, 30 s standing; Vanilla then PsychoV35 | DLSS and NVIDIA FG run; tone-mapper changes recover. Bridge refused Native/DLAA setting requests without changing the setting; DLAA was subsequently tested by a separate cold start |
| 2026-11-12 | d4a69456 + PsychoV35 DLSS port, renderer source 878c957c; host 7afe6ae1; on run 1 | Windows 11 Home, build 26300 | RTX 4070 12 GB, NVIDIA 621.01 | i7-13700F, 32 GB | 2560x1440 output, 2560x1421 window; DLSS Quality 1707x960, K; Native HDR PsychoV35, white 200 / peak 400 nits; sharpness 0; NVIDIA FG 2x; RX_VK_PSYCHOV35_DLSS=1 | the main outdoor area child save, cold start, 30 s standing after settling | RGBA16F input/output accepted, HDR DLSS feature; repeated 600/600 DLSS evaluations and 600/600 NVIDIA FG sets, zero missing-input fallbacks in steady intervals. Method and limits |
| 2026-11-12 | Same renderer and host; off run 1 | Windows 11 Home, build 26300 | RTX 4070 12 GB, NVIDIA 621.01 | i7-13700F, 32 GB | Same scene/settings as on run 1, RX_VK_PSYCHOV35_DLSS=0 | Same copied save, cold start, 30 s standing | No DLSS feature; repeated 0/600 NVIDIA FG sets and 600/600 missing-input fallbacks in steady intervals |
| 2026-11-12 | Same renderer and host; off run 2 | Windows 11 Home, build 26300 | RTX 4070 12 GB, NVIDIA 621.01 | i7-13700F, 32 GB | Same scene/settings as off run 1 | Same copied save, cold start, 30 s standing | Reproduced 0/600 NVIDIA FG sets and 600/600 missing-input fallbacks |
| 2026-11-12 | Same renderer and host; on run 2 | Windows 11 Home, build 26300 | RTX 4070 12 GB, NVIDIA 621.01 | i7-13700F, 32 GB | Same scene/settings as on run 1 | Same copied save, cold start, 30 s standing; then Vanilla/PsychoV35 switches and a camera turn | Reproduced 600/600 DLSS evaluations and NVIDIA FG sets in steady intervals; tone-mapper switching returned to RGBA16F, history reset; actor motion collected |
| 2026-11-12 | feat/auto-preset (the Auto (this PC) branch, on main), game and renderer built from it | Ubuntu 26.04.1, kernel 7.0, a private KWin virtual output 2560x1440 at 60 Hz (headless runs, scripted input) | RTX 4070 Ti 12 GB, 595.99.02 | Ryzen 7 9800X3D, 32 GB | Quality preset Auto (this PC); frame generation off; caps 60 and 90 (HOS_FPS_MAX), and 4K output set by hand for the step-down run |
The test save’s spawn point, then the starting forest and the main outdoor area | First start: tier 4 (High), DLSS Quality, output 2560x1440. Same PC again: kept. RX_VK_AUTO_PRESET_VRAM_MB=6144: Low, DLSS Performance, output 2048x1152; back to 12 GB: High again. A row changed: Custom, and Auto left it. In play: 1440p at 60, the GPU at 8.4 of 16.7 ms: kept; 4K at a 90 cap, frames missed (about 80 fps), the GPU at 10.6 of 11.1 ms: 67% -> 58% -> 50%; 8 GB faked at 60, the GPU at 8.3 ms: Balanced (58%) -> Quality (67%). No validation, device-lost or “not made” line |
| 2026-11-12 | v2.1.0-rc.3-256-g51da046-dirty (source 0212cb04): main with the frame generation maker shown, built from source (llvm-mingw, NVIDIA DLSS SDK with nvngx_dlssg.dll 310.9.16) |
Windows 11 Pro 25H2 (26200), 3440x1440 window | RTX 5060 Ti 16 GB, 617.42 | Core Ultra 7 270K Plus, 32 GB | Upscaler NVIDIA DLSS Quality, Frame generation by Auto, frame rate cap 240, V-Sync Auto | Fire Temple, a save, standing | 2x: the row Auto: DLSS FG, the FPS counter Base 63, DLSS FG at 126-129 FPS (DLSS-G made 600 of 600 presents). Frame generation Auto: Auto: Built-in FG, its line why, Base 71, Built-in FG at 240 FPS. The pause menu and the title screen: Auto: Built-in FG (no DLSS frame). Built-in chosen: Built-in FG. 6x, 4x, 2x, Auto in turn: no fault |
| 2026-11-12 | v2.1.0-rc.3-250-gdea9b646 (source 5821b893): main c544828f + this change (before #2711); against v2.1.0-rc.3-248-gc544828f (source 08d5a9cc), main | Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, default options; nothing else running, RX_VK_TEST_ALLOWED_MB=6900 (an 8 GB card’s allowance; this card cannot page) |
the main outdoor area, a scripted five-minute walk | This change: 58.9 fps, 27 of 29 VRAM lines past the allowance (worst +1,393 MB), no “room in the blocks: not squeezed” line. Main: 58.7 fps, 26 of 28 past (worst +1,641 MB), 26 such lines. The texture budget was already 0 MB on main there; what held the VRAM was the image pool’s blocks made at the start (#2438, fixed since by #2711): main with RX_VK_IMAGE_PREALLOC=0 was past the allowance in 0 of 29 lines, 59.3 fps. No device lost |
| 2026-11-12 | as above | as above | as above | as above | 2560x1440, DLAA, default options; nothing else running, RX_VK_TEST_ALLOWED_MB=6144 (a 6 GB card’s) |
Castle Town to the main outdoor area (an area change) | This change: 51.2 fps, 4 of 6 VRAM lines past the allowance (worst +2,465 MB), no “not squeezed” line. Main: 51.2 fps, 4 of 6 (worst +2,333 MB), 4 such lines; with RX_VK_IMAGE_PREALLOC=0 worst +799 MB. No device lost |
| 2026-11-12 | v2.1.0-rc.3-253-g2fb430bd (source e6840837): main c544828f + #2585 + this change’s first version, whose 8 ms bound was then the default (now RX_VK_TRANSLATE_WAIT_MS=8) |
Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, default options; nothing else running, RX_VK_TEST_ALLOWED_MB=10200; RX_VK_NO_PROGRAM_CACHE=1 (a cold program bank); the 8 ms bound against RX_VK_TRANSLATE_WAIT_MS=-1 (waited out, as main) |
From the title into Castle Town, then out to the main outdoor area | Two runs each. Bound: 50.0 / 47.4 fps, frames over 100 ms 22 / 23, over 250 ms 8 / 6; 1.80 / 1.70 s waiting for shaders; 20,592 / 23,208 draws left out. Waited out: 44.9 / 45.3 fps, 39 / 34, 11 / 11; 4.94 / 4.85 s; 4,404 / 4,368 left out (pipelines still building). Screenshots every 60 frames: nothing missing; the HUD at the first frame after the load was missing waited out, present with the bound. No device lost |
| 2026-11-12 | as above | as above | as above | as above | as above, with the program bank warm | as above | 47.2 fps both ways, no waits for shaders and no draws left out either way |
| 2026-11-12 | v2.1.0-rc.3-259-g31ddd7a5 (source 0a02ef62): main edaef2a9 + this change and other open PRs | Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32; nothing else running, RX_VK_TEST_ALLOWED_MB=10200 |
Castle Town to the main outdoor area, in a folder of the same translator with the stamp file removed | The start dropped the 1,233 MB cache (“1 files made for another translator dropped (1233 MB)”) and the stage libraries were made again; the next start dropped nothing. No device lost |
| 2026-11-12 | v2.1.0-rc.3-154-g2e790fa9 (source 8c169744): main 15646d18 + this change; against v2.1.0-rc.3-153-g15646d18-dirty: main with a temporary measuring patch | Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32; nothing else running, RX_VK_TEST_ALLOWED_MB=10200 |
the main outdoor area, a scripted five-minute walk, in a folder whose pipeline cache had grown to 2,364 MB | The first start dropped the 2,364 MB file; the prepare made a new one of 681 MB. Working set mean 17,708 / 17,550 MB (peak 18,535 / 18,318) against 18,693 MB (peak 19,349) without the change. No pipeline waits, no draws left out, no device lost |
| 2026-11-12 | v2.1.0-rc.3-252-g221410ba (source 39fd6b45): main ee29dd55’s exe, and the same exe with this change’s manifest line patched in (the segment heap) | Windows 11 (26200), no window (unattended, scripted input) | RX 6700 XT 12 GB, AMD 26.10.1 | Ryzen 5 5600, 16 GB | 2560x1440 output, AMD FSR 4 Performance, the game’s caches (warm), RX_VK_NO_PREDICT=1 RX_VK_PRECOMPILE=0 |
the main outdoor area by day, standing, 150 s; four pairs alternated | fps median NT heap / segment heap: 50.8 / 50.1, 51.2 / 50.7, 49.5 / 51.3, 47.3 / 49.2; GPU 13.0-13.6 ms in both; peak private 21.1-21.4 GB in both. No device lost |
| 2026-11-12 | v2.1.0-rc.3-299-gdae19300 (source c7435484): main dae19300 + this change, its own hookshot.exe | Windows 11 (26200), no window (--prepare-shaders) |
RX 6700 XT 12 GB, AMD 26.10.1 (Vulkan driver 0x800161) | Ryzen 5 5600 (6 cores, 12 threads), 16 GB | Cold: empty Hookshot caches, a new folder path (AMD’s driver cache empty for it) | The shader preparation, no game | The segment heap: vk: preparation: 12 workers for the shader programs; first part (vk: prepared ... in) 982 s; 2,153 predicted pipelines in the first 300 s; peak private memory 12.1 GB |
| 2026-11-12 | as above, the same hookshot.exe with the manifest’s heapType line taken out (the NT heap) |
as above | as above | as above | as above | as above | 6 workers ... (half the CPUs); first part 1,478 s; 1,115 predicted pipelines in the first 300 s; peak private memory 9.0 GB |
| 2026-11-12 | v2.1.0 (source e561602e) | Arch Linux, KDE Plasma Wayland, 2560x1440 window (mailbox, udmabuf on) | RX 9070 XT 16 GB, Mesa RADV 26.2.6 | Core i5-13600KF, 32 GB | 2560x1440 output, V-Sync off, parallel recorder on (RX_VK_RECORDER=thread, RX_VK_BATCH=0), RX_VK_MIRROR_MB=4096, idle live-import cap on (RX_VK_IMPORT_HIGH=25, RX_VK_IMPORT_LOW=20: ~3,858 MB high / ~3,086 MB low on the 15,434 MB system heap budget) |
Long session across the game world (176,640 presents, ~1.5 h) | 63.5-65.6 fps at 641-708 draws, 67.2-68.3 fps at 186 draws throughout the session. 257 idle live imports (7,872 MB in all) given back; live imports stayed at 50-64 pools (2,256-2,720 MB), capture submit 0.18-0.43 ms (scan 0.08 ms over 3,923-4,251 MB), queue submit 0.32-0.73 ms, readbacks of GPU results waited 0.64-1.14 ms each. No device lost |
| 2026-11-12 | v2.1.0 (source e561602e) | as above | as above | as above | 2560x1440 output, V-Sync off, default longshot_env.txt (default 2,048 MB mirror budget, default recorder, imports_give_back off on the 15,434 MB system heap budget since budget >= 12 GB) |
Long session across the game world (684,360 presents, ~2 h) | At 10 min (40,920 presents): 60+ fps, 55 live imports (2,000 MB), capture submit 0.27 ms (scan 0.05 ms over 734 ranges, 2,417 MB), queue submit 0.55 ms, readback wait 2.19 ms each (270 of 3,000). After ~2 h (684,120 presents): faded to 44-48.4 fps at 303-408 draws (vi: real-time step on: ratio 1.352, game's frame rate 44); 221 live imports (7,392 MB; 67 retired given back, 2,992 MB), working set 15,351 MB (peak 16,107 MB, 1,511 MB swap), capture submit 0.50 ms (scan 0.09 ms over 334 ranges, 7,218 MB), queue submit 1.11-1.14 ms, readback wait 6.29-7.03 ms each (312-340 of 3,000). No device lost |
| 2026-11-11 | v2.1.0-rc.3-154-g3c89158e (source 1eba3e78): main 15646d18 + this change | Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32; nothing else running, RX_VK_TEST_ALLOWED_MB=10200, RX_VK_FRAME_TRACE=1 |
From the title into Castle Town, then out to the main outdoor area (an area change) | Two runs (3,480 and 3,600 frames traced). The new columns filled in: summed late_wait_ms 260 / 319 ms, readback_wait_ms 13 / 13 ms, tex_frame_ms 323 / 555 ms; late_wait_ms was never over its frame’s memsync_ms (summed 1,671 / 1,808 ms). No device lost |
| 2026-11-12 | v2.1.0-rc.3-120-g1bde1868 (source 62267c81), the native-zstd branch built from source (llvm-mingw, no DLSS) | Windows 11, 5120x1440 (32:9) | RX 9070 XT, Vulkan driver 0x8001a0 | Ryzen 7 5700X3D, 64 GB | HOS_NATIVE_ZSTD=2 |
the main outdoor area, the standard test: the save loaded, the camera turning at full speed, frame rate unlocked (HOS_FPS_MAX=240), 20 s window; 2 warm-ups, then 4 runs an arm, alternating |
fps 43.7-45.0, CPU 111.5-116.3 ms a frame, p99 28.5-30.0 ms, worst 45.0-56.7 ms, no frame over 100 ms; save load 28.0-28.7 s; 3 core escapes |
| 2026-11-12 | v2.1.0-rc.3-120-g1bde1868 (source 62267c81), the native-zstd branch built from source (llvm-mingw, no DLSS) | Windows 11, 5120x1440 (32:9) | RX 9070 XT, Vulkan driver 0x8001a0 | Ryzen 7 5700X3D, 64 GB | HOS_NATIVE_ZSTD=1 |
the main outdoor area, the standard test: the save loaded, the camera turning at full speed, frame rate unlocked (HOS_FPS_MAX=240), 20 s window; 2 warm-ups, then 4 runs an arm, alternating |
fps 42.3-44.8, CPU 112.8-114.9 ms a frame, p99 28.3-36.4 ms, worst 54.4-333.1 ms, 4 frames over 100 ms and 2 over 250, all in the first run; save load 25.9-28.9 s; 4 core escapes |
| 2026-11-12 | v2.1.0-rc.3-120-g1bde1868 (source 62267c81), the native-zstd branch built from source (llvm-mingw, no DLSS) | Windows 11, 5120x1440 (32:9) | RX 9070 XT, Vulkan driver 0x8001a0 | Ryzen 7 5700X3D, 64 GB | Defaults | the main outdoor area, the standard test: the save loaded, the camera turning at full speed, frame rate unlocked (HOS_FPS_MAX=240), 20 s window; 2 warm-ups, then 4 runs an arm, alternating |
fps 44.0-44.9, CPU 112.3-117.9 ms a frame (process cores / fps), p99 28.7-37.6 ms, worst 55.8-220.6 ms, 2 frames over 100 ms in 4 runs, none over 250; save load 26.2-29.2 s; 4 core escapes. Control for HOS_NATIVE_ZSTD below |
| 2026-11-12 | v2.1.0-rc.3-250-ge375837c (main c544828f + this change + a test patch that makes every mirror refused for the budget MIRROR_REFUSED, so that this card has refused pools at all) against v2.1.0-rc.3-249-gcd9e28dd (main + the same test patch) |
Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, RX_VK_MIRROR_MB=256 (most pools refused a mirror); nothing else running |
Castle Town to the main outdoor area (an area change), a five-minute the main outdoor area walk, the starting forest | The “vk: imports:” line at the end. Main: 20 retired pools still imported (1,488 MB) after the area change, 13 (992 MB) in the walk, only 2 ever given back. This change: 0 still imported in all three, 21 (1,472 MB), 15 (1,040 MB) and 11 (512 MB) given back. No device lost, no error |
| 2026-11-12 | v2.1.0-rc.3-154-g2e790fa9 (source 8c169744): main 15646d18 + #2684; RX_VK_IMPORT_CHUNK_MB=4 (this change’s default) against unset (16) |
Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32; nothing else running, RX_VK_TEST_ALLOWED_MB=6144, RX_VK_MIRROR_MB=1536 |
the main outdoor area, a scripted five-minute walk | Two pairs: VRAM peak 5,865 / 5,887 MB with 4 MB chunks against 6,135 / 6,129 MB with 16 (allowance 6,144); the GPU’s longest frame 13.5 / 12.5 ms against 24.1 / 21.7 ms; 53.0 / 52.4 fps against 50.3 / 54.9 (no frame-rate claim). No device lost |
| 2026-11-12 | as above | Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32; nothing else running, RX_VK_TEST_ALLOWED_MB=10200 |
the starting forest, standing, the camera turning | 4 MB chunks: 58.2 fps, mirrors 843 MB, VRAM mean 6,672 MB, longest import 44 ms; 16 MB: 57.0 fps, 1,154 MB, 7,281 MB, 66 ms. No hitch from a pool across a chunk’s end. No device lost |
| 2026-11-12 | v2.1.0-rc.3-153-g15646d18-dirty (source f0a7c2bd): main 15646d18 with a temporary measuring patch (the mirrors’ padding counted); RX_VK_IMPORT_CHUNK_MB=4 against unset (16) |
Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32; nothing else running, RX_VK_TEST_ALLOWED_MB=10200 |
the main outdoor area, a scripted five-minute walk; Castle Town to the main outdoor area (an area change) | Walk: mirrors 1,236-1,268 MB with 4 MB chunks against 1,645-1,657 MB with 16, 58.1 / 58.6 fps against 59.0 / 58.2, imports 91-94 (0.8 s in the driver) against 59-65 (0.7 s). Area change: VRAM mean 7,859 against 8,357 MB, 50.5 against 50.4 fps. No device lost |
| 2026-11-11 | v2.1.0-rc.3-156-g50a6887e (source 44083c56): main 15646d18 + #2683 + this change; against v2.1.0-rc.3-154-g3c89158e (source 1eba3e78): main 15646d18 + #2683 | Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, RX_VK_COMPACT=1, RX_VK_COMPACT_PCT=85, RX_VK_ASTC_BC_ALL=1, RX_VK_IMAGE_BLOCK_MB=32; nothing else running, RX_VK_TEST_ALLOWED_MB=10200 |
From the title into Castle Town (the title-to-game load), then out to the main outdoor area | Two runs each. The frame at the title-to-game load (the movie’s end) took 258 and 259 ms without the change; with it no frame in that stretch was over 60 ms. No device lost |
| 2026-11-12 | v2.1.0-rc.3-260-gb7d175e4f (source 43edcbeb), stock fixed build before rebase, no local counters | Arch Linux, kernel 7.2.15-1-omarchy, headless | Intel UHD 770 (RPL-S), Mesa ANV 26.2.5-arch1.1 | Core i9-13900K, 93 GB reported by the game | 1920x1080, Minimum, FG off, precompile/warm/prediction off, 32 GiB process limit; warmed save/cache; Khronos validation 1.4.363 (RX_VK_VALIDATION=1), duplicate-message cap off |
the starting forest, child, standing; 60 s of confirmed play | VUID-vkCmdDrawIndirect-pStrides-04913: 0 (stock main 51da0464 control: 968 overall, 957 in its 60 s play window). No other errors or new message IDs; existing unused-fragment-output warnings remain. Exit 0, no device loss or GPU fault. Forest/HUD render with this iGPU’s known false colours. No performance claim; full per-ID counts in PR #2796 |
| 2026-11-11 | v2.1.0-rc.3-246-g87991bc8 (source 6d41d308), main built from source (llvm-mingw, no DLSS) | Windows 11, 1280x720 window (mailbox, V-Sync Auto) | RTX 3080 10 GB, 621.14 | Core i9-12900K, 64 GB | Default options, uncapped (HOS_FPS_MAX=1000); a warm-up run first, no pipeline compiled while measuring |
the main outdoor area (adult), by the Lost Woods bridge where the Cheats teleport puts Link, standing; 120 s from 45 s after the teleport | Two runs: 78.4 and 79.3 fps, median frame 12.18 and 12.23 ms, 30 and 37 frames over 20 ms, 1,335-1,453 draws. Readbacks of GPU results: about 1,550 of 3,000 waited, 0.04-0.07 ms each (medians). Against the nightly below for #2602 |
| 2026-11-11 | v2.1.0-rc.3-193-g55b976c0 (source ae9e7c1b), the last nightly before #2602, built from source as above | as above | as above | as above | as above | as above | Two runs, interleaved with the two above: 59.7 and 60.8 fps, median frame 16.39 and 16.32 ms, 264 and 22 frames over 20 ms, 1,355-1,445 draws. Readbacks of GPU results: 172 of 3,000 waited, 5.80 and 5.56 ms each (medians) |
| 2026-11-12 UTC | main 51da0464 / PR2649 84a006d5, matching host and renderer | Windows 11, windowed | RTX 5070 Ti 16 GB, NVIDIA 621.01 | Ryzen 7 9800X3D, 32 GB | 1280x720, 60 fps cap, FG off, precompile/prediction off, RX_VK_PLAY_CACHE=0; identical 1,124,480,456-byte base plus one play file and warmed translator inputs restored before each process | starting-area balcony checkpoint, 30 s after loaded state; eight fresh processes: A/A then AB, BA, AB | All exit 0; checkpoint and final cache hashes match. Paired peak private reductions 924.9 / 1,064.3 / 900.3 MiB, but A/A varied 979.9 MiB. Working-set peak deltas -8.7 / -59.9 / +74.3 MiB; no consistent steady-memory, physical-memory, FPS or startup benefit established |
| 2026-11-12 UTC | PR2649 23c50087 against main 51da0464; reported by dankman42 | Windows 11, windowed | RTX 3080 10 GB, NVIDIA 621.14 | i9-12900K, 64 GB | 1280x720, default options; fixed 1,615,665,731-byte base plus three play files restored each run; one warm-up per build, three interleaved pairs | Adult the village save, teleport to the main outdoor area, 30 s, clean quit; three base saves per measured run | Reported median peak private 26,418 vs 27,831 MB, working set level; all saves published and a further start read the saved cache. Earlier-head external evidence, not a local or current-head rerun; no FPS/Linux/AMD/Intel test |
| 2026-11-12 UTC | v2.1.0-rc.3-234-g7820853d (source 3e8f38f4), matching file-backed-cache host/renderer/image | Windows 11, windowed | RTX 5070 Ti 16 GB, NVIDIA 621.01 | Ryzen 7 9800X3D, 32 GB | 1720x720 requested window, 3440x1440 configured output, game upscaler 67%, 60 fps cap, FG off | starting-area balcony checkpoint, three fresh processes | Loaded checkpoint and rendered scene/HUD; native-menu Quit saved the base cache, 1,124,480,456 to 1,124,543,762 bytes; next process read the saved cache. All exit 0. Functional load/save/read, not a speed or campaign result |
| 2026-11-11 | v2.1.0-rc.3-46-gddafc1f (source c872b22b): the nightly with #2314’s gpu/vk/vk_memory.cpp (97c1a202) swapped in |
Ubuntu 26.04.1, kernel 7.0.0, GNOME 50 Wayland | RTX 4070 12 GB, 615.71.09 | Ryzen 7 5800X3D, 64 GB | 2560x1440 output, DLSS DLAA (model K), frame generation off, Balanced preset (texture budget 6144 MB), Reflex on, RX_VK_PARALLEL=4 |
the starting forest, the stream by the village: fresh launch, then again after about 10 minutes of exploring | Fresh 59.0 fps at about 2,505 draws (memory sync 3.9 ms); after exploring 49.5 fps at about 2,760 draws (memory sync 6.3 ms). 99.96% of the texture-shadow copies came from staging. No device lost |
| 2026-11-11 | as above | as above | as above | as above | as above, RX_VK_SHADOW_STAGE=0 |
as above | Fresh 63.2 fps at about 2,431 draws (memory sync 3.0 ms); after exploring 33.1 fps at 2,431-2,669 draws (memory sync 15.9 ms). No device lost |
| 2026-11-11 | v2.1.0-rc.3-46-gddafc1f (source d13bde75), nightly | as above | as above | as above | as above, without #2314 | the starting forest, the stream by the village, about 2 minutes after launch | 59.2 fps (56.6-60.0 over 29 windows of 120 presents) at about 2,262 draws; replay submit 12.1 ms, memory sync 4.6 ms. No device lost |
| 2026-11-11 | as above | as above | as above | as above | 2560x1440 output, DLSS DLAA (model K), frame generation 3x, texture budget 6144 MB, Reflex on, RX_VK_PARALLEL=4 |
the starting forest, the stream by the village, shortly after launch; then free play | Standing 60.9 fps (59.1-61.4 over 24 windows of 120 presents) at about 2,455 draws. About 8 minutes after launch, in the main outdoor area just north of Lake Hylia by the scarecrows, the device was lost after a 5.4 s GPU memory import: VK_ERROR_DEVICE_LOST, kernel Xid 109 then Xid 31, VRAM 10.4 of 10.8 GB allowed. Log on #1000 |
| 2026-11-11 | host v2.1.0-main-local (source fc2ed0d0); renderer main 49f75dc8 + this branch (FSR 3.1 on the temporal chain) | Gentoo Linux, kernel 7.3.3, KDE Plasma Wayland, windowed | RX 9070 XT 16 GB, Mesa RADV 26.2.6 | Ryzen 9 9950X3D, 32 GB | 2560x1440 output, AMD FSR 3.1 at 100% and Quality (67%), 60 fps cap; RX_VK_FSR3_CHAIN=0 for the comparison runs |
the starting forest (the lost-woods path, a 28 s camera turn), Castle Town market at noon | FSR 3.1 drew every frame on the chain (600 of 600 per status line), no evaluation failures. starting-area, 4 runs each: 59.3 fps on the chain, 59.2 with its own inputs (1% low 48.1 and 46.4). Market, 3 runs each: 52.5 and 53.1 fps (runs vary by about 0.9). Standing still at 67%: edge flicker 1.67 against 3.01, 1-pixel outlines 2 against 12 per megapixel; at 100% the same within noise |
| 2026-11-11 | v2.0.14 (source f2796384) + the value_rec last-part fix (renderer only) |
Anatase 20261005.5 (Fedora-based), Steam Gaming Mode (gamescope, Flatpak Steam), 1280x720 window | ROG Ally X: Ryzen Z1 Extreme iGPU, RADV PHOENIX, Mesa 26.2.3 | Ryzen Z1 Extreme, 24 GB (15.6 GB visible) | 1920x1080 output, game’s upscaler 67%, FG 2x, pacing on; HOS_GUEST_MEMFD=1, RX_VK_VRAM_MB=3072 |
Death Mountain Trail, walking and a fight | 57 s recording: 0 one-frame luminance spikes; 31.3 fps median (the game’s own frames, over 300 draws) |
| 2026-11-11 | host v2.1.0-main-local (source fc2ed0d0); renderer main 49f75dc8 + #1941 + #1942 + #2243 + this change, built with FSR 4 | Gentoo Linux, kernel 7.3.3, KDE Plasma Wayland, windowed | RX 9070 XT 16 GB, Mesa RADV 26.2.6 | Ryzen 9 9950X3D, 32 GB | 2560x1440 HDR10 (HDR Native, paper white 110 nits, peak 400), AMD FSR 4 Native AA through fsr4-vk (FSR 4.1.1, FP8) | the starting forest (the lost-woods path) | The game presented PQ and FSR 4 was made HDR (log: HDR colour, linear); 600 of 600 frames drawn, no device lost. Brightest pixel PQ code 132 (the paper white) with RX_VK_FSR4_HDR=0, 135-145 with the change (two runs); the rest of the scene within one code value, brightness steady frame to frame. SDR unchanged |
| 2026-11-11 | v2.1.0-rc.3-41-g59b9bb75 (main fab8cb3 + the native float add, unmerged then) | Arch Linux, no window (unattended, scripted input) | RTX 4090, 615.78.08 | Core i9-13900K, 96 GB | 2048x1152 output, Resolution 100%, Game speed per frame, RX_VK_PASS_TIMING=2; float adds native (OpAtomicFAddEXT) |
the village (child, day), standing still | GPU time 7.2 ms a frame (median of 56 120-present windows); 59.3 fps mean (60 cap). No device lost |
| 2026-11-11 | as above | as above | as above | as above | as above, RX_VK_NATIVE_FADD=0 (the compare-and-swap loop, as before) |
as above | GPU time 7.8 ms a frame (median of 57); 59.3 fps mean. No device lost |
| 2026-11-11 | as above | as above | as above | as above | as above, float adds native | Death Mountain Trail (adult, night), standing still | GPU time 5.4 ms a frame (median of 57); 59.4 fps mean. No device lost |
| 2026-11-11 | as above | as above | as above | as above | as above, RX_VK_NATIVE_FADD=0 |
as above | GPU time 5.8 ms a frame (median of 57); 59.6 fps mean. No device lost |
| 2026-11-11 | v2.1.0 (source 633eb435) with the 3D BC attachment fix | Windows 10 Enterprise 19045, offscreen | GTX 1080, 581.80 | Core i9-7900X, 64 GB | 1920x1080 offscreen, compute mesh fallback, RX_VK_SKIP_PROG unset |
New game, opening cutscene into the first dungeon and the companion dialogue | Recorded all 128 noise draws. The dialogue rendered and the run reached 15,840 presents without device loss. A standalone Vulkan probe lost the device on a direct BC4 volume attachment clear, then read back all 16,384 blocks correctly through the separate attachment and copy path |
| 2026-11-11 | PR #2357, author report; exact build not recorded | Not recorded | Not recorded | Not recorded | FOV 50% to 150%; also with ultrawide enabled | Gameplay aiming; area not recorded | The PR author reports aiming works across the tested FOV range and with ultrawide mode. Manual gameplay confirmation; no performance measurements reported |
| 2026-11-11 | v2.1.0-rc.3-53-g49f75dc-dirty (source 52a1c006), local feat/adjustable-fov |
Fedora 44, headless | RX 7800 XT 16 GB, RADV Mesa 26.1.8 | Ryzen 7 5700X, 32 GB | 1920x1080, game’s upscaler, FG off, RX_VK_PRECOMPILE=0; copied saves and separate settings |
Dodongo’s Cavern 2F, stationary by the chest | Scripted Gameplay menu changes 100% to 125% and back to 100%. Inspected captures: wider scene at 125%, original framing restored at 100%, HUD placement unchanged; new slider and help fit at 1280x720. Options file saved 125 then 100. No performance comparison; Windows, ultrawide, aiming and cutscene transitions not exercised in that run |
| 2026-11-10 | v2.0.13 (source cfc565a2) + the value_rec last-part fix (renderer only) |
Anatase 20261005.5 (Fedora-based), KDE Plasma desktop, windowed | ROG Ally X: Ryzen Z1 Extreme iGPU, RADV PHOENIX, Mesa 26.2.3 | Ryzen Z1 Extreme, 24 GB (15.6 GB visible) | 1920x1080 output, FSR 3.1 Quality, FG 2x, pacing on, cap 120 | the main outdoor area, walking and fights | 42 s recording: 3 one-frame luminance spikes, 0 of 15/255 or more; 103-137 sync waits in 5 s (1.4-1.8 s); FPS overlay 38-49 |
| 2026-11-10 | v2.0.13 (source cfc565a2) | as above | as above | as above | as above | the main outdoor area, Castle Town | 69 s recording: 104 one-frame luminance spikes, 15 of 15/255 or more (HUD stretched across the screen, grass flat white, whole-screen flashes); 4-109 sync waits in 5 s (31-232 ms); FPS overlay 45-70 |
| 2026-11-11 | v2.1.0-rc.3-58-g06ce5785 (main 49f75dc + the compaction hold’s fix and reserve, and the compaction on by default; unmerged then) | Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, no switch set (the compaction on by default) against RX_VK_COMPACT=0; nothing else running |
the main outdoor area, a scripted five-minute walk (not standing still) | Two runs each, counted from the end of the load. On: 59.8, 59.9 fps; 66/12/2 and 36/2/1 frames over 25/33/50 ms; 1% low 37.7, 40.9 fps; the driver’s count 9.05-9.07 GB in the mean (85% of what it allows), 88-91% at most; the write-back shadows never given back. Off: 56.0, 56.0 fps; 633/60/15 and 408/72/28; 1% low 25.8, 24.1 fps; the count 10.21-10.29 GB in the mean (95-96%), 97-98% at most; the shadows given back 8 times in each. No validation, device-lost or “not made” line |
| 2026-11-11 | as above | as above | as above | as above | 2560x1440, DLAA, RX_VK_TEST_ALLOWED_MB=7000, no other switch against RX_VK_COMPACT=0 |
the main outdoor area, standing, the camera turning for 50 s (not still) | One run each, counted from the end of the load. On: 60.0 fps, 1,203 MB of textures uploaded, 4,498 retired, 1 frame over 25 ms, 1% low 45.4 fps. Off: 59.9 fps, 2,660 MB, 8,332 retired, 8 frames over 25 ms, 1% low 40.7 |
| 2026-11-11 | v2.1.0-rc.3-50-g3acf477a (main 4d68bc8 + the same two commits, rebased; unmerged then) against v2.1.0-rc.3-48-g4d68bc83 (main) | Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, RX_VK_COMPACT=1, RX_VK_TEST_ALLOWED_MB=7000; nothing else running |
the main outdoor area, standing, the camera turning for 50 s (not still) | One run each, counted from the end of the load. The branch: 60.0 fps, 1,187 MB of textures uploaded, 3,445 retired, no frame over 25 ms, 1% low 47.2 fps. Main: 59.6 fps, 6,842 MB, 13,290 retired, 16 frames over 25 ms, 1% low 39.1. While the save loaded, the branch’s count passed the allowance once (7,016 of 7,000 MB) and the write-back shadows were given back once; main’s did not. No validation, device-lost or “not made” line |
| 2026-11-11 | as above | as above | as above | as above | 2560x1440, DLAA, RX_VK_COMPACT=1, RX_VK_ASTC_BC_ALL=1, RX_VK_TEST_ALLOWED_MB=9500 |
the main outdoor area, a scripted five-minute walk (not standing still) | One run each, counted from the end of the load. The branch: 29/8/1 frames over 25/33/50 ms, the write-back shadows never given back, 4,753 MB uploaded, 1% low 42.3 fps. Main: 104/26/6, the shadows given back 5 times, 7,383 MB, 1% low 33.6 |
| 2026-11-11 | v2.0.11-782-gf0c6e73 (main 3932421 + the compaction hold’s fix and reserve, unmerged then) | Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, RX_VK_COMPACT=1, RX_VK_TEST_ALLOWED_MB=7000; nothing else running |
the main outdoor area, standing, the camera turning for 50 s (not still) | Two runs: 60.0, 60.0 fps; 1,284 and 1,284 MB of textures uploaded; 3,814-4,044 textures retired; 7 and 5 frames over 25 ms; 1% low 42.4, 44.3 fps. No validation, device-lost or “not made” line |
| 2026-11-11 | v2.0.11-780-g3932421 (main) | as above | as above | as above | as above | as above | Two runs: 59.0, 59.2 fps; 6,573 and 5,954 MB uploaded; 12,622-13,701 textures retired; 42 and 27 frames over 25 ms; 1% low 34.8, 35.3 fps. Without RX_VK_COMPACT: 60.0 fps, 1,313 MB, 3,481 retired, 6 frames over 25 ms, 1% low 42.5 |
| 2026-11-11 | v2.0.11-782-gf0c6e73 (as above) | as above | as above | as above | 2560x1440, DLAA, RX_VK_COMPACT=1, RX_VK_ASTC_BC_ALL=1, RX_VK_TEST_ALLOWED_MB=9500 |
the main outdoor area, a scripted five-minute walk (not standing still) | Two runs: 86/17/1 and 92/13/1 frames over 25/33/50 ms; the write-back shadows never given back; GPU time over 12.5 ms in 0 of 126 and 2 of 125 windows of 120 presents; 1% low 35.9, 36.9 fps. Without the cap (11.2 GB allowed): 96/19/0 and 65/8/1, shadows never given back, 1% low 35.9, 38.2 |
| 2026-11-11 | v2.0.11-780-g3932421 (main) | as above | as above | as above | as above | as above | Two runs: 120/29/10 and 93/16/3 frames over 25/33/50 ms; the shadows given back 4 and 5 times; GPU time over 12.5 ms in 8 of 124 and 11 of 124 windows; 1% low 30.6, 35.5 fps. Without the cap: 93/24/7 and 97/28/11, shadows given back twice in each, 1% low 33.2, 31.2 |
| 2026-11-11 | host v2.1.0-main-local (source fc2ed0d0); renderer main 49f75dc8 + #1941 + #1942 + #2243 + this change, built with FSR 4 | Gentoo Linux, kernel 7.3.3, KDE Plasma Wayland, windowed | RX 9070 XT 16 GB, Mesa RADV 26.2.6 | Ryzen 9 9950X3D, 32 GB | 2560x1440 HDR10 (HDR Native, paper white 110 nits, peak 400), AMD FSR 4 Native AA through fsr4-vk (FSR 4.1.1, FP8) | the starting forest (the lost-woods path) | The game presented PQ and FSR 4 was made HDR (log: HDR colour, linear); 600 of 600 frames drawn, no device lost. Brightest pixel PQ code 132 (the paper white) with RX_VK_FSR4_HDR=0, 135-145 with the change (two runs); the rest of the scene within one code value, brightness steady frame to frame. SDR unchanged |
| 2026-11-11 | v2.1.0 (source 633eb435) with the 3D BC attachment fix | Windows 10 Enterprise 19045, offscreen | GTX 1080, 581.80 | Core i9-7900X, 64 GB | 1920x1080 offscreen, compute mesh fallback, RX_VK_SKIP_PROG unset |
New game, opening cutscene into the first dungeon and the companion dialogue | Recorded all 128 noise draws. The dialogue rendered and the run reached 15,840 presents without device loss. A standalone Vulkan probe lost the device on a direct BC4 volume attachment clear, then read back all 16,384 blocks correctly through the separate attachment and copy path |
| 2026-11-11 | main d7a73062 + input-origin/game-clock log witnesses; numeric-input prediction candidate renderer (matched API14) |
Windows 11 Home N, 26300 | RX 9070 XT 16 GB, 32.0.32015.12012 | Ryzen 7 5700X3D, 64 GB | 2560x720, 32:9 camera, serial, FG off, wall-clock scripted save load, periodic captures | the main outdoor area, load only, 95 seconds | 4,219 consecutive native trace rows with all seven wall-clock phase fields ordered. All 14 logged game-clock timestamps exactly match native trace endpoints; the input origin is logged. Original frame 2400 shows terrain and HUD. No CPU-time, movement-speedup or clean-validation claim; validation off. First game-clock baseline endpoint is not logged |
| 2026-11-11 | Host and clean renderer fab8cb37; candidate GPU write-back observations on that base |
Windows 11, windowed | RTX 5070 Ti, driver 0x9b404000 |
Ryzen 7 9800X3D, 32 GB | 3840x2160 output, main scene 2880x1620, 1280x720 window, frame generation off, cap 144, exclusive scripted input | starting-area balcony, fresh copy of preserved checkpoint per process | Six 30 s windows in clean/off/on/on/off/clean order: mean FPS 75.249 / 75.523 / 74.242 by mode. Observations on: 95/97 reports, 0 unavailable chunks; changed chunks 15.11/15.03%, covering 13.06/13.00% of compared volume. All six exit normally, original checkpoint unchanged. Two windows per mode; no temperature/clock telemetry, equivalence or culling claim. |
| 2026-11-10 | main 117eae7 (git describe: v2.0.11-730-g117eae7), built with make_game; log: longshot unknown (source 4ca6c9ae) |
NixOS 26.11, Linux 7.2.7; headless Sway 1.12 + Xwayland (unattended, scripted input) | RX 7900 XT (RADV NAVI31), Mesa 26.2.3 | Ryzen 5 7600X, 32 GB | 2048x1152 game output, 1920x1080 window, Auto render resolution, game’s own upscaler, Lanczos window filter, FG 2x, udmabuf on, audio off; mesh mode: RX_VK_FORCE_MESH_FALLBACK=1 only |
the starting forest, the player character’s house balcony (from a save); standing still for the sample, then scripted movement | 139 s total; 46.6 fps median of the last five stationary samples. No GPU hang; stopped by the test |
| 2026-11-10 | as above | as above | as above | as above | as above; mesh mode: RX_VK_TASK_MESH_FALLBACK=1 only |
as above | 139 s total; 43.8 fps median of the last five stationary samples. No GPU hang; stopped by the test |
| 2026-11-10 | as above | as above | as above | as above | as above; mesh mode: native mesh (both fallback switches unset) | as above | 137 s total; 45.4 fps median of the last five stationary samples. No GPU hang; stopped by the test |
| 2026-11-10 | as above | as above | as above | as above | as above; mesh mode: RX_VK_FORCE_MESH_FALLBACK=1 only |
as above | 139 s total; 46.5 fps median of the last five stationary samples. No GPU hang; stopped by the test |
| 2026-11-10 | as above | as above | as above | as above | as above; mesh mode: RX_VK_TASK_MESH_FALLBACK=1 only |
as above | 139 s total; 44.0 fps median of the last five stationary samples. No GPU hang; stopped by the test |
| 2026-11-10 | as above | as above | as above | as above | as above; mesh mode: RX_VK_FORCE_MESH_FALLBACK=1 only |
as above | 139 s total; 46.5 fps median of the last five stationary samples. No GPU hang; stopped by the test |
| 2026-11-10 | as above | as above | as above | as above | as above; mesh mode: RX_VK_TASK_MESH_FALLBACK=1 only |
as above | 138 s total; 43.9 fps median of the last five stationary samples. No GPU hang; stopped by the test |
| 2026-11-10 | as above | as above | as above | as above | as above; mesh mode: native mesh (both fallback switches unset) | as above | 198 s total; 44.9 fps median of the last five stationary samples. No GPU hang; stopped by the test |
| 2026-11-10 | as above | as above | as above | as above | as above; mesh mode: RX_VK_FORCE_MESH_FALLBACK=1 only |
as above | 231 s total; 46.1 fps median of the last five stationary samples. No GPU hang; stopped by the test |
| 2026-11-10 | as above | as above | as above | as above | as above; mesh mode: RX_VK_TASK_MESH_FALLBACK=1 only |
as above | 179 s total; 43.9 fps median of the last five stationary samples. No GPU hang; stopped by the test |
| 2026-11-10 | v2.0.14 (source 3a9998c3), Longshot Builder; later v2.0.14 + the scale-image SAMPLED fix (as in #1940) | Windows 11 Pro 26200, fullscreen | RTX 5090, 621.14 | Ryzen 7 5800X3D, 64 GB | 3840x2160 output, DLSS Quality (67%), Steam Controller (2026) | Castle Town market, the starting forest | Plays well. On the release: HDR Expanded a black picture, Native flat (#1613); with the fix Expanded is right, frame generation 3x too. One crash in nvoglv64.dll (0xc0000005) after about 17,000 frames with frame generation 5x. Steam Controller started from Steam: no gyro until Steam’s SDL_GAMECONTROLLER_IGNORE_DEVICES and SDL_GAMECONTROLLER_ALLOW_STEAM_VIRTUAL_GAMEPAD were cleared (#1897) |
| 2026-11-10 | cc80fde on 9447e1d, Vulkan API 14 |
Windows 11 Pro 26300, windowed | RTX 5070 Ti, 621.01 | Ryzen 7 9800X3D, 32 GB | 3440x1440 game output, fixed FG 2x, parallel replay (2 workers, batch 1), pacing off | the player character’s house in starting-area, copied checkpoint | 6,688 presents logged. Swapchain 1444x720 → 1440x829 → 1444x720, then 1444x720 after minimize/restore. The scene returned and the window closed normally. Surface loss and deferred admission were exercised with injected results in the source fixture. |
| 2026-11-10 | v2.0.11-449-g1987d8c (main caaee0b + the VRAM compaction branch, unmerged then) | Windows 11, no window (unattended, scripted input) | RTX 4070 SUPER 12 GB, 617.42 | Ryzen 9 9950X3D (SMT off), 32 GB | 2560x1440, DLAA, RX_VK_COMPACT=1; other programs open (10.2-10.5 GB of VRAM allowed) |
the main outdoor area, a scripted five-minute walk (not standing still) | Two runs: 58.2, 58.6 fps mean; VRAM 8.3, 7.9 GB; GPU time over 12.5 ms a frame in 6 of 122 and 10 of 123 120-present windows; 88, 62 frames over 33 ms; 1% low 21.1, 26.5 fps. No device lost |
| 2026-11-10 | as above | as above | as above | as above | as above, without the switch | as above | Two runs: 53.2, 54.4 fps mean; VRAM 9.5, 9.6 GB; the write-back shadows given back in 73 of 111 and 80 of 114 windows; GPU time over 12.5 ms in 69 of 111 and 55 of 114; 107, 112 frames over 33 ms; 1% low 19.4, 20.6 fps. No device lost |
| 2026-11-10 | host v2.0.11-438-gcd544a7 (source 8bdd7a54); renderer main 8bac8aa + #1502 + the host-clock capture change |
Windows 11 Home N, 26300 | RX 9070 XT 16 GB, 32.0.32015.12012 | Ryzen 7 5700X3D, 64 GB | 2560x720 window, 32:9 camera, serial, FG off, dense capture, core validation; fresh prepared shader bank | Boot, title, save load; gameplay not reached | 96 s: 1,476 consecutive original capture records; all 1,461 records covered by the flushed native trace align one-to-one with native returns. Movement window rejected: no filled gameplay interval. No new validation IDs; existing 02251, and two 08737 errors during full preparation. No gameplay speed or performance claim |
| 2026-11-10 | v2.0.11-447-gcbe49eb, sampler admission branch (main 8bac8aa + sampler checks + staging fixture repair) | Windows 11, windowed | RTX 5070 Ti, 621.01 | Ryzen 7 9800X3D, 32 GB | 3440x1440 picture, native upscaling, fixed FG 2x, 1 MB recording streams; serial and parallel replay (2 workers, batch 1) | the starting forest, inside the player character’s house; copied checkpoint | Both runs rendered the scene, maximized/restored (client 1444x720 to 1440x829 and back), and exited normally. Six representative PNGs decoded at 3440x1440; original checkpoint hashes unchanged. Healthy-path validation only; no FPS comparison or native OOM recovery claim |
| 2026-11-10 | host 0f5d423a (API13 diagnostic backport), renderer main dadecd53 |
Windows 11 | RX 9070 XT 16 GB, 32.0.32015.12012 | Ryzen 7 5700X3D, 64 GB | 32:9, serial, FG off, GPU profiling/validation off; menu-only CPU program bank plus fixed warmed application pipeline caches; sparse screenshots | the main outdoor area, first entry, control first | 96 s; reached gameplay, terrain checked, no fatal error, original saves unchanged. 557 readiness waits totaling 30.725 s; longest native-present interval 32.568 s. Empty 40-60 s window rejected. Driver/OS caches uncontrolled; no FPS claim |
| 2026-11-10 | same host, renderer d39e988c (#1573) |
as above | as above | as above | identical nine-file starting cache hashes and settings | the main outdoor area, first entry, candidate second | 96 s; terrain/grass/castle checked, no fatal error, original saves unchanged. 815 readiness waits totaling 3.120 s; longest native-present interval 2.114 s. CPU decision tests do not prove GPU writes; no speedup claim |
| 2026-11-10 | same host, renderer main dadecd53 |
as above | as above | as above | identical nine-file starting cache hashes and settings | the main outdoor area, first entry, control repeated third | 96 s; terrain checked, no fatal error, original saves unchanged. 557 readiness waits totaling 0.923 s; longest native-present interval 2.128 s. Unchanged control also lost its long freeze: do not attribute the first/second difference to #1573 |
| 2026-11-10 | main b2bb214 (+ #1447 #1451 #1520 #1660 #1619 #1448 #1654 #1612 #1600, unmerged then) | Ubuntu 26.04.1, KDE Plasma 6.6.6 Wayland | RTX 4070 Ti 12 GB, 595.99.02 | Ryzen 7 9800X3D, 32 GB | 3840x2160 output, DLSS Quality, no mods | the castle road (child) | 48-50 fps. Same spot at 2560x1440 output: 56.3 fps. No device lost, saves written |
| 2026-11-10 | main b2bb214 + #1459 (Dear ImGui menu) | as above | as above | as above | as above, windowed | Title, the castle road | F1 menu, help box, prompts, the game’s font and the FPS badge all correct; 60 fps |
| 2026-11-10 | main b2bb214 | as above | as above | as above | HDR on | start-up | KDE offers an HDR10 surface (vk: HDR: an HDR10 surface (format 58)) |
| 2026-11-09 | main 7eac5bf | as above | as above | as above | 3840x2160, DLSS Quality, no mods | the castle road | 52.8 fps (replay submit 17.5 ms, memory sync 4.8 ms, GPU 12.8 ms) |
| 2026-11-09 | main 7eac5bf | as above | as above | as above | fullscreen (HOS_FULLSCREEN=1) |
the castle road | fullscreen works, no device lost |
| 2026-11-09 | v2.0.11-142 (7dbdcdc) | as above | as above | as above | 3840x2160, DLSS Quality | Water Temple 1F, the castle road, the first dungeon 2F | Water Temple walls missing with the 23.3 GB dump, drawn with stand-in textures (#1308) or a complete dump. Slingshot and bow aiming: no fps change (51-52 fps, #1305). Pause / unpause: no freeze |
Limits of the original RTX 4070 Ti runs
- Real mouse aiming with Mouse camera = Native after https://github.com/NotRexed/Hookshot/pull/1462 (the camera states were traced, but not tested with a real mouse).
- Frame generation at 120 Hz or above (the displays here top out at 75 Hz).
- HDR output on an HDR monitor (the surface is offered, the picture was not judged on one).
