Beanstalk Longshot test farm
Menu

Report rp_24abd5b090f1f8b3 fail

#1866 perf(vk): the recorder takes a whole part and submits it, so the render thread never waits at a part's end

windows+nvidia on Samsheperd/agent-a9f512,

Machine

Commit
f08644e37cd3ce2e6a4801195b6c39d7a2df9f7e
Baseline
89c3d21ea169ed46124e6c97acd90587f660b75a
App
nightly-195-f9b6d598, attempt 1
Game build
e0f3ccfa
os
windows
cpu
AMD Ryzen 9 5900X 12-Core Processor
gpu
NVIDIA GeForce RTX 3080 Ti
tags
windowsnvidianvidia-proprietarydgpu
driver
621.24.0
ram_gb
64
vram_mb
12,084
gpu_vendor
nvidia
os_version
Windows 10.0.26300

Builds

main (baseline)

797 s warm

from 89c3d21ea169

lift
80.7 s
read
0.7 s
finish
29.4 s
compile
679.7 s

PR

36.8 s warm

Program cache: warm

Scenarios

three-areas

ok
Metric (median)mainPRΔ
exit_code00±0%
run_hitch_count1815−16.7%
log_forbidden_hits00±0%
log_expected_missing00±0%
run_frametime_max_ms323.44376.71+16.5%
field-standing.fps_mean108.12108.41+0.3%
kokiri-standing.fps_mean108.87101.17−7.1%
market-standing.fps_mean79.9982.53+3.2%
field-after-turn.fps_mean117.18115.55−1.4%
field-standing.draws_mean1,190.151,197.5+0.6%
field-standing.hitch_count00±0%
kokiri-after-turn.fps_mean109.6798.61−10.1%
kokiri-standing.draws_mean694.6643.6−7.3%
market-after-turn.fps_mean81.1283.55+3%
market-standing.draws_mean2,478.752,440.45−1.5%
field-after-turn.draws_mean732739+1%
kokiri-standing.hitch_count00±0%
market-standing.hitch_count00±0%
field-after-turn.hitch_count00±0%
field-standing.stutter_count00±0%
kokiri-after-turn.draws_mean614.65692.5+12.7%
market-after-turn.draws_mean2,461.452,468.85+0.3%
kokiri-after-turn.hitch_count00±0%
kokiri-standing.stutter_count00±0%
market-after-turn.hitch_count00±0%
market-standing.stutter_count0.50−100%
field-after-turn.stutter_count00±0%
field-standing.frametime_max_ms14.3515.92+10.9%
field-standing.frametime_p50_ms9.219.16−0.5%
field-standing.frametime_p99_ms11.3211.34+0.2%
kokiri-after-turn.stutter_count0.50−100%
market-after-turn.stutter_count01—
kokiri-standing.frametime_max_ms12.9712.63−2.6%
kokiri-standing.frametime_p50_ms9.069.7+7.1%
kokiri-standing.frametime_p99_ms11.1211.4+2.5%
market-standing.frametime_max_ms28.8418.77−34.9%
market-standing.frametime_p50_ms12.4912.05−3.5%
market-standing.frametime_p99_ms15.3914.81−3.8%
field-after-turn.frametime_max_ms11.9814.92+24.6%
field-after-turn.frametime_p50_ms8.528.62+1.2%
field-after-turn.frametime_p99_ms10.2910.94+6.3%
kokiri-after-turn.frametime_max_ms16.8717.68+4.8%
kokiri-after-turn.frametime_p50_ms9.029.96+10.4%
kokiri-after-turn.frametime_p99_ms11.9112.16+2.1%
market-after-turn.frametime_max_ms16.9525.04+47.8%
market-after-turn.frametime_p50_ms12.2511.89−2.9%
market-after-turn.frametime_p99_ms15.1815.67+3.3%
field-standing.frametime_1pct_low_ms12.0712.44+3%
kokiri-standing.frametime_1pct_low_ms11.6211.74+1%
market-standing.frametime_1pct_low_ms17.2515.86−8.1%
field-after-turn.frametime_1pct_low_ms10.7712.87+19.5%
kokiri-after-turn.frametime_1pct_low_ms12.9714+7.9%
market-after-turn.frametime_1pct_low_ms15.8217.76+12.3%
screenshot_diff—0.37—

Pictures against main: kokiri diff 0.6966 (main against itself 0.6969), field diff 0.0482 (main against itself 0.0612), market diff 0.061 (main against itself 0.1136)

Runs (4)
Metricmain 0pr 1pr 2main 3
hangnononono
crashednononono
exit_code0000
run_hitch_count21171315
log_forbidden_hits0000
log_expected_missing0000
run_frametime_max_ms487.32531.87221.54159.56
field-standing.fps_mean108.56108.77108.05107.69
kokiri-standing.fps_mean102.08101.39100.96115.66
market-standing.fps_mean80.383.8581.2279.68
field-after-turn.fps_mean119.33115.97115.12115.03
field-standing.draws_mean1,191.31,198.81,196.21,189
field-standing.hitch_count0000
kokiri-after-turn.fps_mean100.2298.5998.62119.11
kokiri-standing.draws_mean647.6643.3643.9741.6
market-after-turn.fps_mean81.8384.1182.9880.41
market-standing.draws_mean2,502.32,415.72,465.22,455.2
field-after-turn.draws_mean729.2741.2736.8734.8
kokiri-standing.hitch_count0000
market-standing.hitch_count0000
field-after-turn.hitch_count0000
field-standing.stutter_count0000
kokiri-after-turn.draws_mean735.7700.1684.9493.6
market-after-turn.draws_mean2,361.52,303.12,634.62,561.4
kokiri-after-turn.hitch_count0000
kokiri-standing.stutter_count0000
market-after-turn.hitch_count0000
market-standing.stutter_count1000
field-after-turn.stutter_count0000
field-standing.frametime_max_ms14.0415.5816.2514.66
field-standing.frametime_p50_ms9.29.179.169.21
field-standing.frametime_p99_ms11.2211.3111.3811.41
kokiri-after-turn.stutter_count0001
market-after-turn.stutter_count0200
kokiri-standing.frametime_max_ms14.3512.3512.9211.58
kokiri-standing.frametime_p50_ms9.559.79.718.58
kokiri-standing.frametime_p99_ms11.6111.3611.4410.63
market-standing.frametime_max_ms35.0716.2921.2522.62
market-standing.frametime_p50_ms12.4511.8812.2312.52
market-standing.frametime_p99_ms15.1713.9115.715.61
field-after-turn.frametime_max_ms11.661514.8412.3
field-after-turn.frametime_p50_ms8.378.628.638.68
field-after-turn.frametime_p99_ms10.1210.6211.2610.46
kokiri-after-turn.frametime_max_ms15.3818.4316.9418.36
kokiri-after-turn.frametime_p50_ms9.759.969.958.28
kokiri-after-turn.frametime_p99_ms12.3211.9212.4111.51
market-after-turn.frametime_max_ms16.8832.1517.9317.02
market-after-turn.frametime_p50_ms12.1911.7812.0112.32
market-after-turn.frametime_p99_ms14.9416.2315.1115.41
field-standing.frametime_1pct_low_ms12.112.4812.3912.04
kokiri-standing.frametime_1pct_low_ms12.2811.6811.810.97
market-standing.frametime_1pct_low_ms17.414.4917.2217.09
field-after-turn.frametime_1pct_low_ms10.612.7412.9910.94
kokiri-after-turn.frametime_1pct_low_ms13.2814.1713.8312.66
market-after-turn.frametime_1pct_low_ms15.6319.2116.3116.02
screenshot_diff—0.030.7—

Rules

ScenarioCellRuleLimitmainPRΔ %Status
three-areascrashFalse00—pass
three-areaskokiri-standing.fps_mean.min_delta_pct10108.87101.17−7.1%pass
three-areaskokiri-after-turn.fps_mean.min_delta_pct10109.6798.61−10.1%fail
three-areasfield-standing.fps_mean.min_delta_pct10108.12108.41+0.3%pass
three-areasfield-after-turn.fps_mean.min_delta_pct10117.18115.55−1.4%pass
three-areasmarket-standing.fps_mean.min_delta_pct1079.9982.53+3.2%pass
three-areasmarket-after-turn.fps_mean.min_delta_pct1081.1283.55+3%pass
three-areasscreenshot.max_diff0.050.70.7—passover the limit, but main differs from itself by 0.6969 here: no more than 1.5x that is the scene, not the PR
three-areaskokiri-standing.stutter_count.max_delta_pct2000±0%pass
three-areaskokiri-after-turn.stutter_count.max_delta_pct200.50−100%pass
three-areasfield-standing.stutter_count.max_delta_pct2000±0%pass
three-areasfield-after-turn.stutter_count.max_delta_pct2000±0%pass
three-areasmarket-standing.stutter_count.max_delta_pct200.50−100%pass
three-areasmarket-after-turn.stutter_count.max_delta_pct2001—failthe baseline is 0
three-areaskokiri-standing.frametime_p99_ms.max_delta_pct1511.1211.4+2.5%pass
three-areaskokiri-after-turn.frametime_p99_ms.max_delta_pct1511.9112.16+2.1%pass
three-areasfield-standing.frametime_p99_ms.max_delta_pct1511.3211.34+0.2%pass
three-areasfield-after-turn.frametime_p99_ms.max_delta_pct1510.2910.94+6.3%pass
three-areasmarket-standing.frametime_p99_ms.max_delta_pct1515.3914.81−3.8%pass
three-areasmarket-after-turn.frametime_p99_ms.max_delta_pct1515.1815.67+3.3%pass
three-areaslog_forbidden_hits.max000—pass

Log (PR side, end)

[hos sead::MainThread] vk: invalidates since the last readbacks line: 8230 kept their virtual core (nothing there to wait for), 3034 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 8400000 (4644065 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 967 ranges, 3224 MB), 232 pages written, 232 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.005, bytes copied 0.000; 25 kept (84 MB), 11 spare buffers, 6334 texture pages copied (24 MB)
[hos host] vk: program picks: 12389104 kept (4537976 by the last picks' textures, a uniform buffer bound since), 2533 new, 5208362 with other textures (of 17600000)
[hos host] vk: write-backs left out (600 submits): 777, 3811.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 750 dirty pages from their submits' captures, 45783 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 880 dispatches, 35286 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5290 runs, 670.1 MB compared; 81 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 156 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11702 in captures, 7 only in the collection, 38 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 8420000 (4656103 pages from the captures)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 8440000 (4668147 pages from the captures)
[hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #375000: event 0x100d6c8028 mode 0 value 268719 (now 268700) (lr 0x806674b0)
[hos sead::MainThread] nvn: DrawMeshTasksIndirect: 620000 so far (x1 0x117e463050 x2 0 x3 0x117e463040 x4 0x8b040)
[hos host] vk: sparse mirrors: 79, 958 MB with memory of 1500 MB (541 MB spared); blocks given memory later 3915, given back 9293; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 20484; 587 binds (100.9 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 137 taken from the spares, 116 made new; 174 kept as spares, 0 destroyed; 37 spares now, 1032 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 777, 3811.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 963 dirty pages from their submits' captures, 46237 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 900 dispatches, 35684 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5296 runs, 669.0 MB compared; 82 pages read into the shadows; 376 MB of shadows (112 MB of pages known, in 154 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11945 in captures, 2 only in the collection, 36 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 155 live (2240 MB), 0 retired still imported (0 MB), 45 retired given back (1072 MB)
[hos host] vk: a frame: 521 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 0 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.05, present 0.15, draw 5.75, descriptors 2.13, texture lookups 0.08, uploads 0.01, pipelines 0.28, programs 1.00, commands 1.89, dispatch 0.62, memory sync 0.42, queue submit 0.00, slot wait 0.05
[hos host] vk: a frame: mirrors 253 (958 MB): 4.92 MB synced in 953.7 copies, 18.08 MB written back
[hos host] vk: 21720 presents, 83.6 fps; a frame: 2639 draws (0 skipped), 118 dispatches, 140 passes, 143 barriers, 6 submits, 0.1 uploads (0.0 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 37474 pipelines (6529723 ms), 0 programs (0 failed), 2646 textures (4627 MB)
[hos host] vk: host time in GPU resource waits: 0.21 ms a frame (fences 0.03, timelines 0.19, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 372.9 full barriers recorded, 191.1 left out (nothing recorded since the one before), 14.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x8: 3.0; 0x9: 1.0; 0x11: 34.6; 0x12: 76.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 6.6 ms a frame (the longest submission 2.1 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.1 (0.2 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 8460000 (4680190 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 3923 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 8480000 (4692249 pages from the captures)
[hos host] vk: mirror of pool 0x7f82800000 (12 MB) given back: idle, 6048 submissions since its sync
[hos host] vk: program picks: 12543752 kept (4592263 by the last picks' textures, a uniform buffer bound since), 2533 new, 5253714 with other textures (of 17800000)
[hos host] vk: replay reads of guest memory as submitted: 8500000 (4704282 pages from the captures)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 274 KB; 0 made on demand (0 MB kept now); c1 stages: at most 123 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 486 sets (ub 5549, sb 2500, cis 41451, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 798, 3914.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 633 dirty pages from their submits' captures, 46737 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 889 dispatches, 35955 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5405 runs, 686.0 MB compared; 83 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 156 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11647 in captures, 1 only in the collection, 37 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: 1 textures retired (2647 left, 4630 MB): budget 0, drawn over 1, dead targets 0
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 968 ranges, 3212 MB), 231 pages written, 231 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.005, bytes copied 0.000; 25 kept (90 MB), 11 spare buffers, 6341 texture pages copied (24 MB)
[hos host] vk: queues ordered by their events: 126000 submissions stopped at a wait, 125998 went on once it was reached, 0 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 0 (stopped 0.00 ms at most); waits met 376466, not met 0; captures at a CPU event's wait 5561, 0 of them at a wait met on arrival
[hos host] vk: replay reads of guest memory as submitted: 8520000 (4716335 pages from the captures)
[hos ModuleSystemWorker1] vcores: 203238 takes, 37071 waits (mean 59 us, longest 15.2 ms, 0 over HOS_VCORE_LOG_MS), 18544 kicks, 18011 yields, 0 starved grants, 0 escapes in 5 s
[hos sead::MainThread] nvn: CommandBufferSignalEvent #270000: event 0x100d6c8044 mode 0 value 269999 (lr 0x802742f0)
[hos host] vk: replay reads of guest memory as submitted: 8540000 (4728383 pages from the captures)
[hos host] vk: sparse mirrors: 78, 951 MB with memory of 1488 MB (536 MB spared); blocks given memory later 3915, given back 9293; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 20934; 588 binds (101.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 137 taken from the spares, 116 made new; 175 kept as spares, 0 destroyed; 38 spares now, 1044 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 777, 3811.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 888 dirty pages from their submits' captures, 45669 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 880 dispatches, 35278 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5287 runs, 670.1 MB compared; 81 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 157 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11637 in captures, 3 only in the collection, 38 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 155 live (2240 MB), 0 retired still imported (0 MB), 45 retired given back (1072 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 983 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 8285 kept their virtual core (nothing there to wait for), 2979 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos ModuleSystemWorker4] nvn: 2050000 recordings made, 74977 live copies (23.5 MB), 3393 texture copies; 3 command buffers' segments (4.1 MB)
[hos host] vk: replay reads of guest memory as submitted: 8560000 (4740465 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 4322660 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: replay reads of guest memory as submitted: 8580000 (4752499 pages from the captures)
[hos host] vk: program picks: 12698117 kept (4646608 by the last picks' textures, a uniform buffer bound since), 2533 new, 5299349 with other textures (of 18000000)
[hos host] vk: write-backs left out (600 submits): 777, 3811.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 586 dirty pages from their submits' captures, 46456 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 892 dispatches, 35613 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5267 runs, 668.9 MB compared; 83 pages read into the shadows; 376 MB of shadows (112 MB of pages known, in 154 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12024 in captures, 5 only in the collection, 37 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 2408 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos rmbl::VibrationThread] hid: 32768 vibration values, strongest 0.00 since the last count
[hos host] vk: a frame: 521 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 0 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 8.97, present 0.15, draw 5.81, descriptors 2.17, texture lookups 0.08, uploads 0.01, pipelines 0.28, programs 1.01, commands 1.90, dispatch 0.62, memory sync 0.41, queue submit 0.00, slot wait 0.05
[hos host] vk: a frame: mirrors 253 (951 MB): 4.92 MB synced in 953.6 copies, 18.08 MB written back
[hos host] vk: 21840 presents, 84.0 fps; a frame: 2639 draws (0 skipped), 118 dispatches, 140 passes, 143 barriers, 6 submits, 0.1 uploads (0.0 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 37474 pipelines (6529723 ms), 0 programs (0 failed), 2652 textures (4639 MB)
[hos host] vk: host time in GPU resource waits: 0.23 ms a frame (fences 0.03, timelines 0.20, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 372.9 full barriers recorded, 191.0 left out (nothing recorded since the one before), 14.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x8: 3.0; 0x9: 1.0; 0x11: 34.6; 0x12: 76.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 6.7 ms a frame (the longest submission 2.1 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.1 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 8600000 (4764524 pages from the captures)
[hos host] vk: a submission's parts ended: 68569 where the game had ended its render pass, 11431 at 512 draws
[hos host] vk: replay reads of guest memory as submitted: 8620000 (4776585 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 968 ranges, 3214 MB), 238 pages written, 238 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.013; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.007, bytes copied 0.000; 25 kept (134 MB), 9 spare buffers, 6351 texture pages copied (24 MB)
[hos host] vk: sparse mirrors: 78, 951 MB with memory of 1488 MB (536 MB spared); blocks given memory later 3915, given back 9293; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 21384; 588 binds (101.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 137 taken from the spares, 116 made new; 175 kept as spares, 0 destroyed; 38 spares now, 1044 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 798, 3914.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 489 dirty pages from their submits' captures, 46872 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 890 dispatches, 36013 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5453 runs, 691.0 MB compared; 86 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 155 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11675 in captures, 3 only in the collection, 37 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 155 live (2240 MB), 0 retired still imported (0 MB), 45 retired given back (1072 MB)
[hos host] vk: 1 textures retired (2662 left, 4674 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8640000 (4788617 pages from the captures)
[hos host] vk: 2 textures retired (2661 left, 4666 MB): budget 0, drawn over 2, dead targets 0
[hos host] vk: 1 textures retired (2662 left, 4670 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8660000 (4800657 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: program picks: 12852738 kept (4701132 by the last picks' textures, a uniform buffer bound since), 2533 new, 5344728 with other textures (of 18200000)
[hos host] vk: write-backs left out (600 submits): 777, 3811.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 569 dirty pages from their submits' captures, 46037 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 887 dispatches, 35352 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5292 runs, 669.9 MB compared; 79 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 156 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11625 in captures, 5 only in the collection, 37 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 8680000 (4812720 pages from the captures)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #380000: event 0x100d6c8028 mode 0 value 272519 (now 272500) (lr 0x806674b0)
[hos host] vk: replay reads of guest memory as submitted: 8700000 (4824815 pages from the captures)
[hos sead::MainThread] nvn: GPU waits in 5 s: 59 sync waits (26 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 5790 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 274 KB; 0 made on demand (0 MB kept now); c1 stages: at most 123 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 486 sets (ub 5549, sb 2500, cis 41451, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 78, 951 MB with memory of 1488 MB (536 MB spared); blocks given memory later 3915, given back 9293; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 21834; 588 binds (101.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 137 taken from the spares, 116 made new; 175 kept as spares, 0 destroyed; 38 spares now, 1044 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 798, 3914.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 296 dirty pages from their submits' captures, 47545 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 911 dispatches, 36371 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5437 runs, 687.6 MB compared; 79 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 157 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12086 in captures, 5 only in the collection, 37 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 155 live (2240 MB), 0 retired still imported (0 MB), 45 retired given back (1072 MB)
[hos sead::MainThread] vk: readbacks of GPU results: 982 of 3000 waited for their write-back, 0.03 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 8350 kept their virtual core (nothing there to wait for), 2914 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: replay reads of guest memory as submitted: 8720000 (4836850 pages from the captures)
[hos Presentation Thread] vi: 21951 frames presented (83.5 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 418, 15.5-17.8 0, 17.8-22 0, 22-30 0, 30-40 0, over 40 0
[hos host] vk: a frame: 504 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 0 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.06, present 0.21, draw 5.84, descriptors 2.22, texture lookups 0.14, uploads 0.02, pipelines 0.27, programs 1.01, commands 1.90, dispatch 0.62, memory sync 0.42, queue submit 0.00, slot wait 0.05
[hos host] vk: a frame: mirrors 253 (951 MB): 4.92 MB synced in 954.9 copies, 18.13 MB written back
[hos host] vk: 21960 presents, 82.3 fps; a frame: 2602 draws (0 skipped), 119 dispatches, 140 passes, 143 barriers, 6 submits, 0.4 uploads (0.0 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 37474 pipelines (6529723 ms), 0 programs (0 failed), 2667 textures (4670 MB)
[hos host] vk: host time in GPU resource waits: 0.26 ms a frame (fences 0.03, timelines 0.23, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 373.4 full barriers recorded, 191.2 left out (nothing recorded since the one before), 14.5 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x8: 3.0; 0x9: 1.0; 0x11: 34.8; 0x12: 76.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 6.7 ms a frame (the longest submission 2.1 ms); 0.5 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.1 (0.4 MB), ASTC or BC 0.4 (0.7 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 8740000 (4848915 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 967 ranges, 3220 MB), 233 pages written, 233 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.005, bytes copied 0.000; 25 kept (84 MB), 11 spare buffers, 6723 texture pages copied (26 MB)
[hos host] vk: program picks: 13006580 kept (4755619 by the last picks' textures, a uniform buffer bound since), 2533 new, 5390886 with other textures (of 18400000)
[hos host] vk: replay reads of guest memory as submitted: 8760000 (4860972 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 798, 3914.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 299 dirty pages from their submits' captures, 47452 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 903 dispatches, 36294 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5442 runs, 687.6 MB compared; 83 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 157 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12011 in captures, 4 only in the collection, 36 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: replay reads of guest memory as submitted: 8780000 (4873000 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 4396183 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: late pages found at the replay (600 takes): 1 pages in 1 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: replay reads of guest memory as submitted: 8800000 (4885046 pages from the captures)
[hos host] vk: sparse mirrors: 78, 951 MB with memory of 1488 MB (536 MB spared); blocks given memory later 3915, given back 9293; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 22272; 588 binds (101.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 137 taken from the spares, 116 made new; 175 kept as spares, 0 destroyed; 38 spares now, 1044 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 735, 3605.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 550 dirty pages from their submits' captures, 44252 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 846 dispatches, 33882 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5112 runs, 634.6 MB compared; 45 pages read into the shadows; 376 MB of shadows (112 MB of pages known, in 155 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11456 in captures, 2 only in the collection, 36 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 155 live (2240 MB), 0 retired still imported (0 MB), 45 retired given back (1072 MB)
[hos host] vk: render targets drawn over from another start, let go: 7129 so far (29124 MB), 6496 of them (26494 MB) between the passes
[hos host] vk: 4 textures retired (2669 left, 4678 MB): budget 0, drawn over 4, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8820000 (4897109 pages from the captures)
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #275000: event 0x100d6c8010 mode 0 value 275002 (lr 0x80660bcc)
[hos host] vk: replay reads of guest memory as submitted: 8840000 (4909154 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 6784 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: program picks: 13160997 kept (4810060 by the last picks' textures, a uniform buffer bound since), 2533 new, 5436469 with other textures (of 18600000)
[hos host] vk: write-backs left out (600 submits): 756, 3708.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 452 dirty pages from their submits' captures, 44554 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 836 dispatches, 34188 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5196 runs, 650.6 MB compared; 48 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 155 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11158 in captures, 6 only in the collection, 35 only by the replay itself
[hos host] vk:   read live, by pool:
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 967 ranges, 3220 MB), 234 pages written, 234 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.010; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (84 MB), 11 spare buffers, 6728 texture pages copied (26 MB)
[hos host] vk: replay reads of guest memory as submitted: 8860000 (4921182 pages from the captures)
[hos host] vk: a frame: 518 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 0 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.04, present 0.15, draw 5.86, descriptors 2.19, texture lookups 0.09, uploads 0.00, pipelines 0.28, programs 1.03, commands 1.92, dispatch 0.62, memory sync 0.40, queue submit 0.00, slot wait 0.05
[hos host] vk: a frame: mirrors 253 (951 MB): 4.94 MB synced in 957.3 copies, 18.13 MB written back
[hos host] vk: 22080 presents, 83.4 fps; a frame: 2636 draws (0 skipped), 121 dispatches, 140 passes, 143 barriers, 6 submits, 0.0 uploads (0.0 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 37474 pipelines (6529723 ms), 0 programs (0 failed), 2679 textures (4717 MB)
[hos host] vk: host time in GPU resource waits: 0.21 ms a frame (fences 0.03, timelines 0.19, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 378.8 full barriers recorded, 191.9 left out (nothing recorded since the one before), 14.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x8: 3.0; 0x9: 1.0; 0x11: 34.6; 0x12: 76.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 6.7 ms a frame (the longest submission 2.1 ms); 0.1 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.1 (0.5 MB), ASTC or BC 0.0 (0.0 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 8880000 (4933231 pages from the captures)
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 981 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker4] vk: invalidates since the last readbacks line: 8618 kept their virtual core (nothing there to wait for), 3158 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: sparse mirrors: 78, 951 MB with memory of 1488 MB (536 MB spared); blocks given memory later 3915, given back 9293; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 22698; 588 binds (101.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 137 taken from the spares, 116 made new; 175 kept as spares, 0 destroyed; 38 spares now, 1044 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 735, 3605.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 407 dirty pages from their submits' captures, 43873 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 838 dispatches, 33505 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5133 runs, 637.5 MB compared; 41 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 156 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11164 in captures, 5 only in the collection, 36 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 155 live (2240 MB), 0 retired still imported (0 MB), 45 retired given back (1072 MB)
[hos host] vk: 1 textures retired (2678 left, 4713 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: 4 textures retired (2679 left, 4717 MB): budget 0, drawn over 4, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8900000 (4945266 pages from the captures)
[hos host] vk: 4 textures retired (2675 left, 4702 MB): budget 0, drawn over 4, dead targets 0
[hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #385000: event 0x100d6c8028 mode 0 value 276319 (now 276300) (lr 0x806674b0)
[hos host] vk: replay reads of guest memory as submitted: 8920000 (4957329 pages from the captures)
[hos host] vk: 3 textures retired (2677 left, 4691 MB): budget 0, drawn over 3, dead targets 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 274 KB; 0 made on demand (0 MB kept now); c1 stages: at most 123 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 486 sets (ub 5549, sb 2500, cis 41451, si 0, utb 0, stb 0)
[hos host] vk: write-backs left out (600 submits): 756, 3708.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 897 dirty pages from their submits' captures, 44755 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 860 dispatches, 34592 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5232 runs, 651.3 MB compared; 57 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 155 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11456 in captures, 3 only in the collection, 35 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: program picks: 13315747 kept (4864264 by the last picks' textures, a uniform buffer bound since), 2533 new, 5481719 with other textures (of 18800000)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 2 textures retired (2682 left, 4685 MB): budget 0, drawn over 2, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8940000 (4969403 pages from the captures)
[hos host] vk: 1 textures retired (2682 left, 4682 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 8960000 (4981489 pages from the captures)
[hos host] vk: sparse mirrors: 78, 951 MB with memory of 1488 MB (536 MB spared); blocks given memory later 3915, given back 9293; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 23136; 588 binds (101.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 137 taken from the spares, 116 made new; 175 kept as spares, 0 destroyed; 38 spares now, 1044 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 777, 3811.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 811 dirty pages from their submits' captures, 45898 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 884 dispatches, 35393 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5327 runs, 670.7 MB compared; 81 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 157 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11672 in captures, 3 only in the collection, 38 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 155 live (2240 MB), 0 retired still imported (0 MB), 45 retired given back (1072 MB)
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 952 ranges, 3108 MB), 230 pages written, 230 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.010; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.004, bytes copied 0.000; 25 kept (84 MB), 11 spare buffers, 6707 texture pages copied (26 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 2874 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 8980000 (4993588 pages from the captures)
[hos host] vk: 1 textures retired (2684 left, 4678 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 9000000 (5005619 pages from the captures)
[hos ModuleSystemWorker1] nvn: 2100000 recordings made, 74947 live copies (24.8 MB), 3409 texture copies; 4 command buffers' segments (4.3 MB)
[hos host] vk: readback pages refreshed from guest memory: 4469249 pages so far (0 held: may be written ahead in their submission)
[hos PreActorModelCompose] vcores: 202266 takes, 36899 waits (mean 57 us, longest 11.1 ms, 0 over HOS_VCORE_LOG_MS), 18327 kicks, 17755 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.05
[hos host] vk: write-backs left out (600 submits): 777, 3811.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 697 dirty pages from their submits' captures, 46535 from earlier ones', 0 from guest memory as it was at the replay; 24 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 893 dispatches, 35682 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5309 runs, 669.4 MB compared; 76 pages read into the shadows; 376 MB of shadows (112 MB of pages known, in 155 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12076 in captures, 5 only in the collection, 37 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: program picks: 13470463 kept (4918656 by the last picks' textures, a uniform buffer bound since), 2533 new, 5527003 with other textures (of 19000000)
[hos host] vk: memory: working set 23425 MB (peak 23543), private 32332 MB, staging 48 MB in 3 arenas, 37 made so far (5 ms)
[hos host] vk: memory heaps: heap 0 (device): 226 blocks 7336 MB, 18650 allocations 6259 MB; used 8331 of 11316 MB; heap 1 (system): 14 blocks 928 MB, 39 allocations 840 MB; used 3277 of 48295 MB;
[hos host] vk: image blocks: 1 blocks 32 MB, 5 images 16 MB; 1 free ranges, the largest 15 MB
[hos host] vk: image blocks (render targets): 29 blocks 928 MB, 164 images 346 MB; 94 free ranges, the largest 32 MB
[hos host] vk: image blocks (textures, compacted): 104 blocks 3328 MB, 2509 images 3100 MB; 1361 free ranges, the largest 17 MB
[hos host] vk: render target images: 4184 reused from the pool, 4072 made new, 3902 destroyed from it; it holds 0 (0 MB)
[hos host] vk: VRAM: 8331 MB used of 11316 MB allowed (1076 MB of it free in the blocks); textures 4678 MB (budget 7049 MB), 2684 textures
[hos host] vk: budget: VRAM 7336 of 11316 MB (8331 MB by the driver, 995 MB unaccounted): textures 4678, blocks free 1076, mirrors 951, shadows 376, other buffers 142, other images 63, staging 48; RAM 3114 MB (the process's private memory 32332 MB): imports 2240, capture spares 746, captures 90, capture table 34, other buffers 4, others 0
[hos host] vk: textures: decoded ASTC 2425 MB, rendered 1063 MB (54 drawn in one frame, kept: 53778 KB), other 810 MB; used in the last 8 presents 2190 MB
[hos host] vk: decoded ASTC used again: 421 textures (1098 MB) copied from ones the game finalized, not decoded; 421 of those retired, 63 listed; 1254 MB of blocks hashed in 143 ms
[hos host] vk:   288 MB: 0x1147c16000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk:   214 MB: 0x10a5b22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 5623 presents ago
[hos host] vk:   214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 5623 presents ago
[hos host] vk:   64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk:   64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk:   54 MB: 0x100030a4200 3072x3072x1 x3 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk:   54 MB: 0x10002d97a00 3072x3072x1 x3 levels 1 NVN 0x32 rendered, last used 1107 presents ago
[hos host] vk:   54 MB: 0x10002fb3c00 3072x3072x1 x3 levels 1 NVN 0x32 rendered, last used 7526 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 73 (19599 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 50x57x1 NVN 0x25: 1 (11 KB); 62x41x1 NVN 0x25: 1 (9 KB); 74x32x1 NVN 0x25: 1 (9 KB); 90x34x1 NVN 0x25: 1 (11 KB); 114x121x1 NVN 0x1: 1 (13 KB); 126x105x1 NVN 0x1: 1 (12 KB); 138x96x1 NVN 0x1: 1 (12 KB); 140x42x1 NVN 0x25: 1 (22 KB); 154x98x1 NVN 0x1: 1 (14 KB); 165x112x1 NVN 0x1: 1 (18 KB); 198x40x1 NVN 0x25: 2 (61 KB); 204x106x1 NVN 0x1: 1 (21 KB); 242x40x1 NVN 0x25: 3 (113 KB); 244x44x1 NVN 0x25: 1 (41 KB); 260x260x1 NVN 0xd: 36 (6334 KB);
[hos host] vk: 2 textures retired (2682 left, 4670 MB): budget 0, drawn over 2, dead targets 0
[hos Presentation Thread] vk: presents after the CPU's signal (600): 78 waited for it, 0.39 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 17.7 ms; VRAM 8.1 of 11.1 GB
[hos host] vk: a frame: 521 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 0 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.12, present 0.20, draw 5.77, descriptors 2.18, texture lookups 0.13, uploads 0.02, pipelines 0.27, programs 0.99, commands 1.88, dispatch 0.62, memory sync 0.40, queue submit 0.00, slot wait 0.05
[hos host] vk: a frame: mirrors 253 (951 MB): 4.94 MB synced in 956.7 copies, 18.11 MB written back
[hos host] vk: 22200 presents, 83.3 fps; a frame: 2639 draws (0 skipped), 120 dispatches, 140 passes, 143 barriers, 6 submits, 0.3 uploads (0.0 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 37474 pipelines (6529723 ms), 0 programs (0 failed), 2682 textures (4670 MB)
[hos host] vk: host time in GPU resource waits: 0.21 ms a frame (fences 0.03, timelines 0.18, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 376.0 full barriers recorded, 191.3 left out (nothing recorded since the one before), 14.6 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x8: 3.0; 0x9: 1.0; 0x11: 34.6; 0x12: 76.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 6.7 ms a frame (the longest submission 2.1 ms); 0.4 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.2 MB), ASTC or BC 0.3 (0.3 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.02 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the render thread (600 presents, 12.00 ms a frame): its jobs 9.42 ms a frame (submits 7.91, presents 0.18, others 1.33), idle 2.58; the game waited for room at 0 presents, 0.00 ms each (0.00 a frame; 1.16 jobs queued at a present); 0 catch-ups (a finish, a texture read back, a write-back not recorded yet: 0), 0.00 ms each; 0 presents held for the CPU's signal, 0.00 ms each (0 let go by RX_VK_HOLD_MS); 90 captures at a CPU signal; the submits' waits read in 0.058 ms a frame; readbacks: 708 next to a texture's storage (no catch-up), 0 in one, 0 read exactly (after the work handed over: among the catch-ups); a present made 8.57 ms after its hand-over (12.95 at most)
[hos host] vk: frame overlap (600 presents): the GPU between recordings, ms a frame, idle + in the next one's uploads: up to a frame's first recording with a draw 1.78 + 0.15 (the window's copy is in the idle), after it 2.87 + 0.52; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.02 ms of GPU time), after its last that reads it 1.00 (0.02 ms)
[hos host] vk: frame ahead (600 presents): 10.26 parts a frame recorded in the signal's job, 76.4% of them took nothing that came with the signal; of the others, 2.42 a frame took pages of the signal's capture into a mirror (86.2 pages a frame) and 0.00 pages read as guest memory had them (0.0 pages a frame); a frame's first part took nothing in 9 of 90 frames, its first three in 9
[hos host] vk: frame ahead, what came with the signal: 90 signals with a capture, each of 131.6 pages in 8.8 pools; between a frame's last submit and that signal 0.00 texture and sampler registrations, 0.00 releases and 0.00 other CPU signals a signal (0 of them with any)
[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 1.15+0.16, 0.16+0.03, 0.31+0.03, 0.78+0.13, 0.15+0.04, 0.16+0.03, 0.04+0.02, 0.03+0.02, 0.05+0.03, 0.02+0.02, 0.01+0.00, 0.00+0.00
[hos host] vk: replay reads of guest memory as submitted: 9020000 (5017662 pages from the captures)
[hos host] vk: 1 textures retired (2681 left, 4666 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: recorder: 13714679 packets run in 10.0 s (busy 2.92 s, spun idle 0.17 s, slept 15084 times, 3720 of them under 200 us), 0 calls made directly (no room); 26868 batches over 834 frames (32.2 a frame, 510 packets a batch), 13506 parts ended and submitted by it; the render thread (the replay lock's holder) waited for it 58 times, 6.95 ms (0.1 times, 0.01 ms a frame); every thread 75 times, 8.02 ms in all, the longest 0.29 ms; vkDestroyImageView 38/6.85 ms (max 0.29), the queue's lock 16/0.87 ms (max 0.19), a part's submit 1/0.20 ms (max 0.20), vkQueueBindSparse 10/0.05 ms (max 0.01), vmaFreeMemory 10/0.05 ms (max 0.01)
[hos host] vk: recorder: of its busy 2.92 s, 0.57 s in parts' ends and submits (the longest 2.01 ms), 0.00 s of that blocked on the queue's lock (0 times, the longest 0.00 ms), so 0.57 s in the ends, submits and holds; 2.35 s running the other packets; 16444 packets a frame
[hos host] vk: 6 textures retired (2675 left, 4643 MB): budget 0, drawn over 6, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 9040000 (5029725 pages from the captures)
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 987 of 3000 waited for their write-back, 0.02 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos ModuleSystemWorker2] vk: invalidates since the last readbacks line: 8054 kept their virtual core (nothing there to wait for), 2954 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 1 textures retired (2674 left, 4639 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: sparse mirrors: 78, 951 MB with memory of 1488 MB (536 MB spared); blocks given memory later 3915, given back 9293; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 23586; 588 binds (101.0 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 137 taken from the spares, 116 made new; 175 kept as spares, 0 destroyed; 38 spares now, 1044 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 798, 3914.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 540 dirty pages from their submits' captures, 46725 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 886 dispatches, 35922 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5436 runs, 686.3 MB compared; 103 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 155 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11543 in captures, 6 only in the collection, 37 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 155 live (2240 MB), 0 retired still imported (0 MB), 45 retired given back (1072 MB)
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos host] vk: 1 textures retired (2676 left, 4647 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 9060000 (5041772 pages from the captures)
[hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 640000 so far (x1 0x117b427030 x2 0 x3 0x117b427000 x4 0xf33080)
[hos host] vk: replay reads of guest memory as submitted: 9080000 (5053800 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 949 ranges, 3096 MB), 245 pages written, 245 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.018; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.012, bytes copied 0.000; 25 kept (86 MB), 11 spare buffers, 6373 texture pages copied (24 MB)
[hos host] vk: 3 textures retired (2673 left, 4635 MB): budget 0, drawn over 3, dead targets 0
[hos host] vk: write-backs left out (600 submits): 777, 3811.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 685 dirty pages from their submits' captures, 45856 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 884 dispatches, 35244 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5287 runs, 670.1 MB compared; 83 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 156 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11692 in captures, 5 only in the collection, 37 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 12746 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: program picks: 13624951 kept (4972934 by the last picks' textures, a uniform buffer bound since), 2533 new, 5572515 with other textures (of 19200000)
[hos host] vk: replay reads of guest memory as submitted: 9100000 (5065823 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 9120000 (5077900 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.05
[hos ModuleSystemWorker2] nvn: CommandBufferSignalEvent #280000: event 0x100d6c8000 mode 0 value 279995 (lr 0x80291440)
[hos host] vk: replay reads of guest memory as submitted: 9140000 (5089923 pages from the captures)
[hos host] vk: 1 textures retired (2678 left, 4632 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: the recording streams: 48 of 1 MB, the most one held 274 KB; 0 made on demand (0 MB kept now); c1 stages: at most 123 a recording; 0 descriptor pools made when a recording's own ran out (0 slots' pools made whole); a recording's pool took at most 486 sets (ub 5549, sb 2500, cis 41451, si 0, utb 0, stb 0)
[hos host] vk: sparse mirrors: 79, 958 MB with memory of 1500 MB (541 MB spared); blocks given memory later 3915, given back 9293; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 24030; 589 binds (101.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 138 taken from the spares, 116 made new; 175 kept as spares, 0 destroyed; 37 spares now, 1032 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 777, 3811.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 888 dirty pages from their submits' captures, 45827 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 883 dispatches, 35390 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5388 runs, 678.6 MB compared; 80 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 157 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11581 in captures, 2 only in the collection, 37 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 155 live (2240 MB), 0 retired still imported (0 MB), 45 retired given back (1072 MB)
[hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #390000: event 0x100d6c8028 mode 0 value 280119 (now 280100) (lr 0x806674b0)
[hos host] vk: a frame: 522 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 0 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 9.18, present 0.18, draw 5.90, descriptors 2.24, texture lookups 0.13, uploads 0.03, pipelines 0.28, programs 1.01, commands 1.91, dispatch 0.63, memory sync 0.46, queue submit 0.00, slot wait 0.05
[hos host] vk: a frame: mirrors 254 (958 MB): 4.98 MB synced in 953.1 copies, 18.22 MB written back
[hos host] vk: 22320 presents, 82.6 fps; a frame: 2643 draws (0 skipped), 118 dispatches, 140 passes, 143 barriers, 6 submits, 0.3 uploads (0.0 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 37474 pipelines (6529723 ms), 0 programs (0 failed), 2678 textures (4632 MB)
[hos host] vk: host time in GPU resource waits: 0.33 ms a frame (fences 0.03, timelines 0.31, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 374.2 full barriers recorded, 191.1 left out (nothing recorded since the one before), 14.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x8: 3.0; 0x9: 1.0; 0x11: 34.7; 0x12: 76.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 6.7 ms a frame (the longest submission 2.1 ms); 0.3 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.1 MB), ASTC or BC 0.3 (0.6 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.01 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 9160000 (5102000 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 9180000 (5114066 pages from the captures)
[hos host] vk: program picks: 13779769 kept (5027432 by the last picks' textures, a uniform buffer bound since), 2533 new, 5617697 with other textures (of 19400000)
[hos host] vk: write-backs left out (600 submits): 798, 3914.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 461 dirty pages from their submits' captures, 47469 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 906 dispatches, 36290 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5728 runs, 716.1 MB compared; 82 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 155 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12049 in captures, 4 only in the collection, 36 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: late pages found at the replay (600 takes): 0 pages in 0 takes (RX_VK_LATE_AT_REPLAY=0: only the ones the capture found)
[hos sead::MainThread] vk: capture: a submit 0.12 ms (lock 0.00, scan 0.05 over 929 ranges, 2986 MB), 234 pages written, 234 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.011; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.005, bytes copied 0.000; 25 kept (94 MB), 10 spare buffers, 5027 texture pages copied (19 MB)
[hos sead::MainThread] nvn: GPU waits in 5 s: 60 sync waits (23 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: 1 textures retired (2682 left, 4633 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 9200000 (5126135 pages from the captures)
[hos sead::MainThread] vk: readbacks of GPU results: 984 of 3000 waited for their write-back, 0.05 ms each; 0 looked at the stopped queues first, 0 waited for their submission's last part (guest memory read in place after theirs), 0 waited for their whole submission (their pages written in place, no mirror)
[hos sead::MainThread] vk: invalidates since the last readbacks line: 8348 kept their virtual core (nothing there to wait for), 2916 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos Presentation Thread] vi: 22367 frames presented (83.1 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 415, 15.5-17.8 1, 17.8-22 0, 22-30 0, 30-40 0, over 40 0
[hos host] vk: readback pages refreshed from guest memory: 4542688 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: replay reads of guest memory as submitted: 9220000 (5138153 pages from the captures)
[hos host] vk: sparse mirrors: 79, 958 MB with memory of 1500 MB (541 MB spared); blocks given memory later 3915, given back 9293; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 24480; 589 binds (101.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 138 taken from the spares, 116 made new; 175 kept as spares, 0 destroyed; 37 spares now, 1032 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 777, 3811.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 675 dirty pages from their submits' captures, 45951 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 898 dispatches, 35318 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5348 runs, 676.8 MB compared; 85 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 157 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 11671 in captures, 2 only in the collection, 37 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 155 live (2240 MB), 0 retired still imported (0 MB), 45 retired given back (1072 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 9469 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: replay reads of guest memory as submitted: 9240000 (5150218 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06
[hos host] vk: 1 textures retired (2705 left, 4632 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 9260000 (5162279 pages from the captures)
[hos host] vk: program picks: 13933924 kept (5081984 by the last picks' textures, a uniform buffer bound since), 2533 new, 5663542 with other textures (of 19600000)
[hos host] vk: write-backs left out (600 submits): 798, 3914.3 MB; 31 bindings followed, 43 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 412 dirty pages from their submits' captures, 47354 from earlier ones', 0 from guest memory as it was at the replay; 25 captures kept; 0 pages the GPU wrote copied over (0 live); 0 kept for the GPU (RX_VK_PROTECT_BARE)
[hos host] vk: mirror syncs' pass (600 submits): 900 dispatches, 36244 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 5418 runs, 688.0 MB compared; 72 pages read into the shadows; 376 MB of shadows (113 MB of pages known, in 156 MB of 1 MB parts); pages the shadows knew that the CPU wrote: 12037 in captures, 6 only in the collection, 37 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 9280000 (5174340 pages from the captures)
[hos host] vk: a frame: 507 draws kept their neighbour's descriptors (0 of them through other slots' registrations)
[hos host] vk: a frame: 0.0 draws left out (not ready), 0.00 waits for shaders or pipelines (0.0 ms); translation 0 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 8.99, present 0.20, draw 5.84, descriptors 2.22, texture lookups 0.11, uploads 0.03, pipelines 0.27, programs 1.01, commands 1.89, dispatch 0.62, memory sync 0.42, queue submit 0.00, slot wait 0.05
[hos host] vk: a frame: mirrors 255 (959 MB): 4.93 MB synced in 953.7 copies, 18.35 MB written back
[hos host] vk: 22440 presents, 82.8 fps; a frame: 2617 draws (0 skipped), 119 dispatches, 140 passes, 143 barriers, 6 submits, 0.5 uploads (0.0 MB), 82 ub updates (57 overlays), GPU-bound wait 0.0 ms (recording slots only); 37474 pipelines (6529723 ms), 0 programs (0 failed), 2705 textures (4632 MB)
[hos host] vk: host time in GPU resource waits: 0.36 ms a frame (fences 0.03, timelines 0.34, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 373.2 full barriers recorded, 191.2 left out (nothing recorded since the one before), 14.7 pass starts ordered in their fragment stages only (draw passes after draw passes)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x8: 3.0; 0x9: 1.0; 0x11: 34.7; 0x12: 76.5; 0x13: 2.0; 0x21: 4.0; 0x28: 16.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 6.7 ms a frame (the longest submission 2.1 ms); 0.5 textures made a frame: colour targets 0.0 (0.0 MB), depth 0.0 (0.0 MB), ASTC or BC 0.5 (0.6 MB), other 0.0 (0.0 MB)
[hos host] vk: a frame: 0.03 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: replay reads of guest memory as submitted: 9300000 (5186420 pages from the captures)
[hos host] control: quit asked by a client
[hos host] vk: indexed descriptor arrays (close): 2661 in 2661 of 52288 program stages, 0 of them near a limit
[hos host] vk: indexed descriptor arrays (close), fragment stages: 2661 texture (17-64: 2661); the longest 64
[hos host] vk: page table census (close): no pool outside its windows, 64-72 GB and 508-512 GB (4258 pools in the table now)
[hos host] control: exiting
[hos host] stop: close work done (a file the game was writing waited for, at most 3 s)