Beanstalk Longshot test farm
Menu

Report rp_6cee7735e34d343e error

#1681 perf(vk): render targets take a pooled image's memory by size, and the pool holds longer on integrated GPUs

windows+nvidia on Flashman4/agent-de36c5,

This report disagrees with its peers.

field-turn-and-stand run_frametime_max_ms: -14.2% against peers' +28.3%
field-turn-and-stand run_hitch_count: +40.0% against peers' +100.0%
field-turn-and-stand#1 run_frametime_max_ms: +4.4% against peers' +67.5%
field-turn-and-stand#2 run_frametime_max_ms: +291.0% against peers' -43.6%
field-turn-and-stand#2 run_hitch_count: +62.5% against peers' +100.0%
field-turn-and-stand#2 value_pool_mb: -100.0% against peers' +357.0%
no scenario could be judged: a step ran out of time at step 2 (wait map HyruleField_Child): timeout: not met in time (then the game was in play on KokiriForest_Child)

Machine

Commit
742b053ae6b0a34dca75a574db424c559f22850f
Baseline
068fb8ebe232d455ef48e817702b03ccc62d8520
App
nightly-206-d6e717d1, attempt 3
Game build
e0f3ccfa
os
windows
cpu
AMD Ryzen 9 5900X 12-Core Processor
gpu
NVIDIA GeForce RTX 3090
tags
windowsnvidianvidia-proprietarydgpu
driver
621.1.0
ram_gb
128
vram_mb
24,322
gpu_vendor
nvidia
os_version
Windows 10.0.26200

Builds

main (baseline)

0 s warm

from 068fb8ebe232

the base was built already

PR

36 s warm

Program cache: warm

Scenarios

field-turn-and-stand RX_VK_TARGET_REUSE_MEMORY=0 (PR only)

ok
Metric (median)mainPRΔ
exit_code00±0%
value_pool_mb388344−11.3%
run_hitch_count57+40%
standing.fps_mean60.0159.97−0.1%
log_forbidden_hits00±0%
standing.draws_mean939.9946.3+0.7%
log_expected_missing00±0%
run_frametime_max_ms235.54202.11−14.2%
standing.hitch_count00±0%
standing.vram_mean_mb9,3019,377.3+0.8%
standing.vram_peak_mb9,3019,420+1.3%
standing.stutter_count00±0%
standing.frametime_max_ms20.8123.33+12.1%
standing.frametime_p50_ms16.7316.67−0.4%
standing.frametime_p99_ms18.7319.29+2.9%
standing.frametime_1pct_low_ms19.3220.83+7.8%
screenshot_diff—0.09—
value_reused_by_size—0—

Pictures against main: field-before diff 0.0074 (main against itself ), field-after-turn diff 0.0468 (main against itself ), field-standing diff 0.0925 (main against itself )

Runs (2)
Metricmain 0pr 1
hangnono
crashednono
exit_code00
value_pool_mb388344
run_hitch_count57
standing.fps_mean60.0159.97
log_forbidden_hits00
standing.draws_mean939.9946.3
log_expected_missing00
run_frametime_max_ms235.54202.11
standing.hitch_count00
value_reused_by_size—0
standing.vram_mean_mb9,3019,377.3
standing.vram_peak_mb9,3019,420
standing.stutter_count00
standing.frametime_max_ms20.8123.33
standing.frametime_p50_ms16.7316.67
standing.frametime_p99_ms18.7319.29
standing.frametime_1pct_low_ms19.3220.83
screenshot_diff—0.09

field-turn-and-stand RX_VK_TARGET_REUSE_MEMORY=1 (PR only)

timeout

a step ran out of time at step 2 (wait map HyruleField_Child): timeout: not met in time (then the game was in play on KokiriForest_Child)

Metric (median)mainPRΔ
value_pool_mb00±0%
run_hitch_count86−25%
log_forbidden_hits00±0%
log_expected_missing00±0%
run_frametime_max_ms274.42286.53+4.4%
value_reused_by_size—6—
Runs (2)
Metricmain 0pr 1
hangnono
crashednono
exit_code——
value_pool_mb00
run_hitch_count86
log_forbidden_hits00
log_expected_missing00
run_frametime_max_ms274.42286.53
value_reused_by_size—6

field-turn-and-stand RX_VK_TARGET_REUSE_MEMORY=2 (PR only)

timeout

a step ran out of time at step 2 (wait map HyruleField_Child): timeout: not met in time (then the game was in play on KokiriForest_Child)

Metric (median)mainPRΔ
exit_code0——
value_pool_mb1380−100%
run_hitch_count813+62.5%
standing.fps_mean59.98——
log_forbidden_hits00±0%
standing.draws_mean949.9——
log_expected_missing00±0%
run_frametime_max_ms220.68862.77+291%
standing.hitch_count0——
standing.vram_mean_mb9,988.5——
standing.vram_peak_mb10,040——
standing.stutter_count0——
standing.frametime_max_ms20.49——
standing.frametime_p50_ms16.69——
standing.frametime_p99_ms18.98——
standing.frametime_1pct_low_ms19.45——
value_reused_by_size—13—
Runs (2)
Metricmain 0pr 1
hangnono
crashednono
exit_code0—
value_pool_mb1380
run_hitch_count813
standing.fps_mean59.98—
log_forbidden_hits00
standing.draws_mean949.9—
log_expected_missing00
run_frametime_max_ms220.68862.77
standing.hitch_count0—
value_reused_by_size—13
standing.vram_mean_mb9,988.5—
standing.vram_peak_mb10,040—
standing.stutter_count0—
standing.frametime_max_ms20.49—
standing.frametime_p50_ms16.69—
standing.frametime_p99_ms18.98—
standing.frametime_1pct_low_ms19.45—

Rules

ScenarioCellRuleLimitmainPRΔ %Status
field-turn-and-standRX_VK_TARGET_REUSE_MEMORY=0 (PR only)crashFalse00—pass
field-turn-and-standRX_VK_TARGET_REUSE_MEMORY=0 (PR only)screenshot.max_diff0.15—0.09—pass
field-turn-and-standRX_VK_TARGET_REUSE_MEMORY=0 (PR only)log_forbidden_hits.max000—pass
field-turn-and-standRX_VK_TARGET_REUSE_MEMORY=1 (PR only)crashFalse00—pass
field-turn-and-standRX_VK_TARGET_REUSE_MEMORY=1 (PR only)screenshot.max_diff0.15———skippedthe plan has no screenshot probe
field-turn-and-standRX_VK_TARGET_REUSE_MEMORY=1 (PR only)log_forbidden_hits.max000—pass
field-turn-and-standRX_VK_TARGET_REUSE_MEMORY=2 (PR only)crashFalse00—pass
field-turn-and-standRX_VK_TARGET_REUSE_MEMORY=2 (PR only)screenshot.max_diff0.15———skippedthe plan has no screenshot probe
field-turn-and-standRX_VK_TARGET_REUSE_MEMORY=2 (PR only)log_forbidden_hits.max000—pass
field-turn-and-standRX_VK_TARGET_REUSE_MEMORY=1 (PR only)standing.frametime_p99_ms: RX_VK_TARGET_REUSE_MEMORY=1 vs RX_VK_TARGET_REUSE_MEMORY=0 (PR)2019.29——skippedthe metric was not measured in both cells
field-turn-and-standRX_VK_TARGET_REUSE_MEMORY=2 (PR only)standing.frametime_p99_ms: RX_VK_TARGET_REUSE_MEMORY=2 vs RX_VK_TARGET_REUSE_MEMORY=0 (PR)2019.29——skippedthe metric was not measured in both cells

Pictures of the failure

The last frame of a step that got stuck, or main's and the PR's picture where they differed more than the plan allows; shrunk, and kept a day: until .

field-turn-and-stand-c1-baseline0-last-frame.png
field-turn-and-stand-c1-baseline0-last-frame.png
field-turn-and-stand-c1-pr1-last-frame.png
field-turn-and-stand-c1-pr1-last-frame.png

Log (PR side, end)

[hos host] vk: frame overlap, by place: idle + uploads before a frame's 2nd to 12th recording with a draw, then all later ones: 0.54+0.14, 0.44+0.12, 0.02+0.03, 0.02+0.03, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[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: queues ordered by their events: 24000 submissions stopped at a wait, 23990 went on once it was reached, 8 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 2 (stopped 50.00 ms at most); waits met 68492, not met 0; captures at a CPU event's wait 1527, 0 of them at a wait met on arrival
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1321 of 3000 waited for their write-back, 0.07 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: 4150 kept their virtual core (nothing there to wait for), 3018 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: 1320000 (671232 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 1310, 5366.5 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 324 dirty pages from their submits' captures, 67290 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): 1316 dispatches, 52436 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6616 runs, 898.2 MB compared; 131 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 23445 in captures, 1 only in the collection, 230 only by the replay itself
[hos host] vk:   read live, by pool:
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.06 over 1280 ranges, 1769 MB), 210 pages written, 210 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (228 MB), 10 spare buffers, 0 texture pages copied (0 MB)
[hos BaseProcExecutor3] vcores: 146428 takes, 19929 waits (mean 60 us, longest 5.6 ms, 0 over HOS_VCORE_LOG_MS), 9652 kicks, 9659 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 1340000 (681404 pages from the captures)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #70000: event 0x10103c8048 mode 0 value 49647 (now 49628) (lr 0x806661f8)
[hos host] vk: readback pages refreshed from guest memory: 663609 pages so far (0 held: may be written ahead in their submission)
[hos ModuleSystemWorker5] nvn: 450000 recordings made, 44296 live copies (12.6 MB), 319 texture copies; 5 command buffers' segments (4.3 MB)
[hos host] vk: recorder: 4024851 packets run in 10.0 s (busy 1.32 s, spun idle 2.87 s, slept 4394 times, 1338 of them under 200 us), 0 calls made directly (no room); the replay waited for it 19988 times, 106.68 ms in all, the longest 0.40 ms; vkEndCommandBuffer 6400/45.86 ms (max 0.40), replay_mu unlock 7218/35.10 ms (max 0.35), vkResetDescriptorPool 6367/25.68 ms (max 0.24), vkDestroyImageView 3/0.05 ms (max 0.02)
[hos host] vk: sparse mirrors: 55, 787 MB with memory of 992 MB (204 MB spared); blocks given memory later 90, given back 1228; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 84 binds (11.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 7 taken from the spares, 63 made new; 15 kept as spares, 0 destroyed; 8 spares now, 156 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1289, 5276.2 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 260 dirty pages from their submits' captures, 66519 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): 1298 dispatches, 51595 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6720 runs, 914.1 MB compared; 140 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 23027 in captures, 5 only in the collection, 234 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 64 live (1092 MB), 0 retired still imported (0 MB), 8 retired given back (172 MB)
[hos ModuleSystemWorker1] nvn: CommandBufferSignalEvent #50000: event 0x10103c801c mode 0 value 49997 (lr 0x802911b8)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06
[hos host] vk: replay reads of guest memory as submitted: 1360000 (691947 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[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 frame: 51 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 6.48, present 0.24, draw 3.51, descriptors 1.40, texture lookups 0.07, uploads 0.00, pipelines 0.15, programs 0.68, commands 1.02, dispatch 0.75, memory sync 0.43, queue submit 0.44, slot wait 0.05
[hos host] vk: a frame: mirrors 70 (787 MB): 4.65 MB synced in 919.8 copies, 16.04 MB written back
[hos host] vk: 4320 presents, 60.0 fps; a frame: 822 draws (0 skipped), 104 dispatches, 142 passes, 144 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 37474 pipelines (7236569 ms), 0 programs (0 failed), 1239 textures (3519 MB)
[hos host] vk: host time in GPU resource waits: 0.51 ms a frame (fences 0.03, timelines 0.48, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 351.2 full barriers recorded, 184.6 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (RX_VK_PASS_BARRIER=game)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x8: 3.0; 0x9: 1.0; 0x11: 45.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 14.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 8.0 ms a frame (the longest submission 4.4 ms); 0.0 textures made a frame
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: write-backs left out (600 submits): 1288, 5274.3 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 264 dirty pages from their submits' captures, 66194 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): 1295 dispatches, 51429 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6629 runs, 898.2 MB compared; 122 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 23145 in captures, 5 only in the collection, 238 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 1380000 (702339 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.06 over 1280 ranges, 1769 MB), 208 pages written, 208 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (212 MB), 12 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: program picks: 1496165 kept (609183 by the last picks' textures, a uniform buffer bound since), 782 new, 903052 with other textures (of 2400000)
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1323 of 3000 waited for their write-back, 0.06 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 ModuleSystemWorker3] vk: invalidates since the last readbacks line: 4058 kept their virtual core (nothing there to wait for), 3110 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: 1400000 (712713 pages from the captures)
[hos Presentation Thread] vi: 4386 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 1, 15.5-17.8 299, 17.8-22 1, 22-30 0, 30-40 0, over 40 0
[hos host] vk: the recording streams: 48 of 2 MB, the most one held 346 KB; 0 made on demand (0 MB kept now); c1 stages: at most 114 a recording; 0 descriptor pools made when a recording's own ran out
[hos host] vk: sparse mirrors: 55, 787 MB with memory of 992 MB (204 MB spared); blocks given memory later 90, given back 1228; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 84 binds (11.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 7 taken from the spares, 63 made new; 15 kept as spares, 0 destroyed; 8 spares now, 156 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1288, 5274.3 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 250 dirty pages from their submits' captures, 66167 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): 1295 dispatches, 51405 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6625 runs, 898.2 MB compared; 142 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 23047 in captures, 7 only in the collection, 234 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 64 live (1092 MB), 0 retired still imported (0 MB), 8 retired given back (172 MB)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07
[hos host] vk: replay reads of guest memory as submitted: 1420000 (723113 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[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 frame: 51 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 6.68, present 0.25, draw 3.63, descriptors 1.48, texture lookups 0.07, uploads 0.00, pipelines 0.15, programs 0.71, commands 1.04, dispatch 0.76, memory sync 0.44, queue submit 0.50, slot wait 0.05
[hos host] vk: a frame: mirrors 70 (787 MB): 4.63 MB synced in 918.2 copies, 16.04 MB written back
[hos host] vk: 4440 presents, 60.0 fps; a frame: 824 draws (0 skipped), 104 dispatches, 142 passes, 144 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 37474 pipelines (7236569 ms), 0 programs (0 failed), 1240 textures (3524 MB)
[hos host] vk: host time in GPU resource waits: 0.58 ms a frame (fences 0.03, timelines 0.55, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 351.4 full barriers recorded, 184.6 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (RX_VK_PASS_BARRIER=game)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x8: 3.0; 0x9: 1.0; 0x11: 45.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 14.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 8.0 ms a frame (the longest submission 4.2 ms); 0.0 textures made a frame
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.06 over 1280 ranges, 1769 MB), 208 pages written, 208 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (236 MB), 11 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: write-backs left out (600 submits): 1288, 5274.3 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 263 dirty pages from their submits' captures, 66637 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): 1308 dispatches, 51763 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6640 runs, 898.2 MB compared; 121 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 23077 in captures, 5 only in the collection, 229 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 1440000 (733562 pages from the captures)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #75000: event 0x10103c8048 mode 0 value 53447 (now 53428) (lr 0x806661f8)
[hos rmbl::VibrationThread] hid: 12288 vibration values, strongest 0.00 since the last count
[hos host] vk: replay reads of guest memory as submitted: 1460000 (743947 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 ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 120000 so far (x1 0x11867f0000 x2 0 x3 0x11867f0000 x4 0x8b040)
[hos host] vk: sparse mirrors: 55, 787 MB with memory of 992 MB (204 MB spared); blocks given memory later 90, given back 1228; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 84 binds (11.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 7 taken from the spares, 63 made new; 15 kept as spares, 0 destroyed; 8 spares now, 156 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1311, 5368.5 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 323 dirty pages from their submits' captures, 66777 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): 1305 dispatches, 52104 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6740 runs, 914.2 MB compared; 117 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 23543 in captures, 4 only in the collection, 235 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 64 live (1092 MB), 0 retired still imported (0 MB), 8 retired given back (172 MB)
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1326 of 3000 waited for their write-back, 0.07 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: 3944 kept their virtual core (nothing there to wait for), 2712 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: 1480000 (754347 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[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 Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.06 over 1280 ranges, 1769 MB), 210 pages written, 210 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (220 MB), 11 spare buffers, 0 texture pages copied (0 MB)
[hos ModuleSystemWorker5] nvn: CommandBufferSignalEvent #55000: event 0x10103c8030 mode 0 value 54999 (lr 0x806612a8)
[hos host] vk: readback pages refreshed from guest memory: 737735 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: a frame: 51 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 6.22, present 0.24, draw 3.40, descriptors 1.38, texture lookups 0.07, uploads 0.00, pipelines 0.15, programs 0.65, commands 0.97, dispatch 0.72, memory sync 0.41, queue submit 0.45, slot wait 0.04
[hos host] vk: a frame: mirrors 70 (787 MB): 4.63 MB synced in 919.5 copies, 16.04 MB written back
[hos host] vk: 4560 presents, 60.0 fps; a frame: 822 draws (0 skipped), 104 dispatches, 142 passes, 144 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 37474 pipelines (7236569 ms), 0 programs (0 failed), 1240 textures (3524 MB)
[hos host] vk: host time in GPU resource waits: 0.58 ms a frame (fences 0.03, timelines 0.55, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 351.3 full barriers recorded, 184.6 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (RX_VK_PASS_BARRIER=game)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x8: 3.0; 0x9: 1.0; 0x11: 45.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 14.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 7.7 ms a frame (the longest submission 3.9 ms); 0.0 textures made a frame
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos rmbl::VibrationThread] vcores: 146809 takes, 19901 waits (mean 58 us, longest 6.1 ms, 0 over HOS_VCORE_LOG_MS), 9707 kicks, 9664 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: write-backs left out (600 submits): 1288, 5274.3 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 295 dirty pages from their submits' captures, 66097 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): 1295 dispatches, 51505 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6630 runs, 898.2 MB compared; 124 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 23089 in captures, 2 only in the collection, 236 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 1500000 (764751 pages from the captures)
[hos host] vk: program picks: 1620849 kept (660654 by the last picks' textures, a uniform buffer bound since), 782 new, 978368 with other textures (of 2600000)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06
[hos host] vk: replay reads of guest memory as submitted: 1520000 (775072 pages from the captures)
[hos host] vk: sparse mirrors: 55, 787 MB with memory of 992 MB (204 MB spared); blocks given memory later 90, given back 1228; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 84 binds (11.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 7 taken from the spares, 63 made new; 15 kept as spares, 0 destroyed; 8 spares now, 156 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1288, 5274.3 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 248 dirty pages from their submits' captures, 66463 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): 1296 dispatches, 51614 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6629 runs, 898.2 MB compared; 122 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 23136 in captures, 3 only in the collection, 234 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 64 live (1092 MB), 0 retired still imported (0 MB), 8 retired given back (172 MB)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.06 over 1279 ranges, 1767 MB), 208 pages written, 208 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 24 kept (204 MB), 12 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[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: 1540000 (785247 pages from the captures)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #80000: event 0x10103c8048 mode 0 value 57247 (now 57228) (lr 0x806661f8)
[hos ModuleSystemWorker4] nvn: 500000 recordings made, 44316 live copies (12.8 MB), 319 texture copies; 5 command buffers' segments (4.3 MB)
[hos host] vk: the recording streams: 48 of 2 MB, the most one held 346 KB; 0 made on demand (0 MB kept now); c1 stages: at most 114 a recording; 0 descriptor pools made when a recording's own ran out
[hos host] vk: write-backs left out (600 submits): 1288, 5274.3 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 280 dirty pages from their submits' captures, 66543 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): 1307 dispatches, 51668 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6620 runs, 898.2 MB compared; 125 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 23054 in captures, 5 only in the collection, 229 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: a frame: 51 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 6.47, present 0.25, draw 3.42, descriptors 1.36, texture lookups 0.07, uploads 0.00, pipelines 0.14, programs 0.65, commands 1.02, dispatch 0.73, memory sync 0.42, queue submit 0.48, slot wait 0.05
[hos host] vk: a frame: mirrors 70 (787 MB): 4.64 MB synced in 919.5 copies, 16.04 MB written back
[hos host] vk: 4680 presents, 60.0 fps; a frame: 822 draws (0 skipped), 104 dispatches, 142 passes, 144 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 37474 pipelines (7236569 ms), 0 programs (0 failed), 1241 textures (3529 MB)
[hos host] vk: host time in GPU resource waits: 0.57 ms a frame (fences 0.03, timelines 0.54, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 351.4 full barriers recorded, 184.6 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (RX_VK_PASS_BARRIER=game)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x8: 3.0; 0x9: 1.0; 0x11: 45.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 14.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 7.9 ms a frame (the longest submission 4.2 ms); 0.0 textures made a frame
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos Presentation Thread] vi: 4686 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 0, 15.5-17.8 300, 17.8-22 0, 22-30 0, 30-40 0, over 40 0
[hos sead::MainThread] vk: readbacks of GPU results: 1329 of 3000 waited for their write-back, 0.06 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: 4113 kept their virtual core (nothing there to wait for), 3311 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: 1560000 (795431 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 (1240 left, 3524 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: replay reads of guest memory as submitted: 1580000 (805664 pages from the captures)
[hos host] vk: sparse mirrors: 55, 787 MB with memory of 992 MB (204 MB spared); blocks given memory later 90, given back 1228; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 84 binds (11.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 7 taken from the spares, 63 made new; 15 kept as spares, 0 destroyed; 8 spares now, 156 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1311, 5368.5 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 321 dirty pages from their submits' captures, 66788 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): 1306 dispatches, 52054 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6731 runs, 914.2 MB compared; 118 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 23550 in captures, 5 only in the collection, 235 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 64 live (1092 MB), 0 retired still imported (0 MB), 8 retired given back (172 MB)
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.06 over 1279 ranges, 1767 MB), 208 pages written, 208 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (234 MB), 11 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 0 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[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: 1600000 (815844 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 1288, 5274.3 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 225 dirty pages from their submits' captures, 66168 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): 1295 dispatches, 51466 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6629 runs, 898.2 MB compared; 130 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 23025 in captures, 4 only in the collection, 237 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: 1 textures retired (1239 left, 3519 MB): budget 0, drawn over 1, dead targets 0
[hos host] vk: memory: working set 20908 MB (peak 20944), private 28577 MB, staging 32 MB in 2 arenas, 25 made so far (5 ms)
[hos host] vk: memory heaps: heap 0 (device): 168 blocks 5310 MB, 14077 allocations 4825 MB; used 6364 of 23554 MB; heap 1 (system): 26 blocks 768 MB, 38 allocations 659 MB; used 1969 of 104029 MB;
[hos host] vk: image blocks: 1 blocks 32 MB, 4 images 12 MB; 1 free ranges, the largest 19 MB
[hos host] vk: image blocks (render targets): 23 blocks 736 MB, 98 images 424 MB; 36 free ranges, the largest 32 MB
[hos host] vk: image blocks (textures, compacted): 58 blocks 1856 MB, 1130 images 1836 MB; 475 free ranges, the largest 12 MB
[hos host] vk: render target images: 58 reused from the pool, 260 made new, 158 destroyed from it, 13 reused by size (on a pooled one's memory, RX_VK_TARGET_REUSE_MEMORY=2), 1 missed after their image was taken by size; it holds 0 (0 MB)
[hos host] vk: render targets not reused by size: 202 with the pool empty (0 of them with targets still in the dead queue), 18 with none in it a present yet, 0 with none of a kind they can use, 8 with all smaller (the largest 32% short on average), 7 with all bigger (over 1.5 times), 25 with some of each, 0 over a block, 0 refused by the driver, 0 not bound; the pool let go of 0 over its bound and 158 idle
[hos host] vk: VRAM: 6364 MB used of 23554 MB allowed (484 MB of it free in the blocks); textures 3519 MB (budget 11000 MB), 1239 textures
[hos host] vk: budget: VRAM 5310 of 23554 MB (6364 MB by the driver, 1054 MB unaccounted): textures 3519, mirrors 787, blocks free 484, shadows 304, other buffers 154, staging 32, other images 28; RAM 1779 MB (the process's private memory 28577 MB): imports 1092, capture spares 416, captures 234, capture table 28, other buffers 9, others 0
[hos host] vk: textures: decoded ASTC 1885 MB, rendered 1023 MB (101 drawn in one frame, kept: 62267 KB), other 344 MB; used in the last 8 presents 2385 MB
[hos host] vk: decoded ASTC used again: 0 textures (0 MB) copied from ones the game finalized, not decoded; 0 of those retired, 3 listed; 378 MB of blocks hashed in 36 ms
[hos host] vk:   288 MB: 0x1150c06000 12288x12288x1 x1 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk:   214 MB: 0x10aea22c00 1365x1365x1 x90 levels 11 NVN 0x80 ASTC, last used 1 presents ago
[hos host] vk:   214 MB: 0x10ab1aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk:   64 MB: 0x114045c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk:   64 MB: 0x1145b99000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk:   64 MB: 0x114045c000 4096x4096x1 x1 levels 1 NVN 0xa, last used 3050 presents ago
[hos host] vk:   54 MB: 0x1000558d800 3072x3072x1 x3 levels 1 NVN 0x32 rendered, last used 3031 presents ago
[hos host] vk:   54 MB: 0x10006146200 3072x3072x1 x3 levels 1 NVN 0x32 rendered, last used 1 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 2 (177 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 260x260x1 NVN 0xd: 1 (175 KB);
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.06
[hos Presentation Thread] vk: presents after the CPU's signal (600): 21 waited for it, 0.11 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 21.2 ms; VRAM 6.2 of 23.0 GB
[hos host] vk: a frame: 51 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 6.45, present 0.25, draw 3.49, descriptors 1.42, texture lookups 0.08, uploads 0.00, pipelines 0.14, programs 0.66, commands 1.02, dispatch 0.74, memory sync 0.42, queue submit 0.46, slot wait 0.05
[hos host] vk: a frame: mirrors 70 (787 MB): 4.63 MB synced in 919.0 copies, 16.04 MB written back
[hos host] vk: 4800 presents, 60.0 fps; a frame: 822 draws (0 skipped), 104 dispatches, 142 passes, 144 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 37474 pipelines (7236569 ms), 0 programs (0 failed), 1239 textures (3519 MB)
[hos host] vk: host time in GPU resource waits: 0.47 ms a frame (fences 0.03, timelines 0.44, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 351.3 full barriers recorded, 184.6 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (RX_VK_PASS_BARRIER=game)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x8: 3.0; 0x9: 1.0; 0x11: 45.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 14.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 8.0 ms a frame (the longest submission 4.6 ms); 0.0 textures made a frame
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: the render thread (600 presents, 16.67 ms a frame): its jobs 7.36 ms a frame (submits 6.71, presents 0.26, others 0.39), idle 9.31; the game waited for room at 0 presents, 0.00 ms each (0.00 a frame; 2.45 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); 27 captures at a CPU signal; the submits' waits read in 0.027 ms a frame; readbacks: 2446 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 6.27 ms after its hand-over (10.79 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 7.05 + 0.32 (the window's copy is in the idle), after it 1.03 + 0.33; after a frame's last recording that reads or writes the game's memory come 1.00 recordings (0.03 ms of GPU time), after its last that reads it 1.00 (0.03 ms)
[hos host] vk: frame ahead (600 presents): 4.78 parts a frame recorded in the signal's job, 89.1% of them took nothing that came with the signal; of the others, 0.52 a frame took pages of the signal's capture into a mirror (3.9 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 18 of 27 frames, its first three in 17
[hos host] vk: frame ahead, what came with the signal: 27 signals with a capture, each of 10.0 pages in 2.1 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: 0.56+0.14, 0.43+0.12, 0.02+0.03, 0.02+0.03, 0.01+0.01, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00, 0.00+0.00
[hos host] vk: replay reads of guest memory as submitted: 1620000 (825971 pages from the captures)
[hos host] vk: program picks: 1745482 kept (712074 by the last picks' textures, a uniform buffer bound since), 782 new, 1053735 with other textures (of 2800000)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #60000: event 0x10103c803c mode 0 value 59999 (lr 0x80661bfc)
[hos host] vk: readback pages refreshed from guest memory: 811438 pages so far (0 held: may be written ahead in their submission)
[hos sead::MainThread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.06 over 1279 ranges, 1767 MB), 210 pages written, 210 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (234 MB), 11 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: replay reads of guest memory as submitted: 1640000 (836176 pages from the captures)
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1314 of 3000 waited for their write-back, 0.06 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 ModuleSystemWorker1] vk: invalidates since the last readbacks line: 3801 kept their virtual core (nothing there to wait for), 2855 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: sparse mirrors: 55, 787 MB with memory of 992 MB (204 MB spared); blocks given memory later 90, given back 1228; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 84 binds (11.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 7 taken from the spares, 63 made new; 15 kept as spares, 0 destroyed; 8 spares now, 156 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1288, 5274.3 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 347 dirty pages from their submits' captures, 66551 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): 1307 dispatches, 51768 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6620 runs, 898.2 MB compared; 125 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 23094 in captures, 4 only in the collection, 234 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 64 live (1092 MB), 0 retired still imported (0 MB), 8 retired given back (172 MB)
[hos nn.npln.Shared] vcores: 147391 takes, 20112 waits (mean 60 us, longest 6.1 ms, 0 over HOS_VCORE_LOG_MS), 9755 kicks, 9742 yields, 0 starved grants, 0 escapes in 5 s
[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] nvn: CommandBufferWaitEvent #85000: event 0x10103c8048 mode 0 value 61047 (now 61028) (lr 0x806661f8)
[hos host] vk: recorder: 4031799 packets run in 10.0 s (busy 1.35 s, spun idle 2.89 s, slept 4398 times, 1353 of them under 200 us), 0 calls made directly (no room); the replay waited for it 20003 times, 108.96 ms in all, the longest 0.34 ms; vkEndCommandBuffer 6405/47.41 ms (max 0.32), replay_mu unlock 7231/35.49 ms (max 0.34), vkResetDescriptorPool 6367/26.06 ms (max 0.29)
[hos host] vk: replay reads of guest memory as submitted: 1660000 (846602 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: write-backs left out (600 submits): 1310, 5366.5 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 229 dirty pages from their submits' captures, 67121 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): 1305 dispatches, 52147 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6614 runs, 898.2 MB compared; 142 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 23383 in captures, 4 only in the collection, 230 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 1680000 (857029 pages from the captures)
[hos host] vk: a frame: 51 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 6.24, present 0.24, draw 3.36, descriptors 1.36, texture lookups 0.07, uploads 0.00, pipelines 0.14, programs 0.66, commands 0.96, dispatch 0.72, memory sync 0.43, queue submit 0.45, slot wait 0.05
[hos host] vk: a frame: mirrors 70 (787 MB): 4.64 MB synced in 919.6 copies, 16.04 MB written back
[hos host] vk: 4920 presents, 60.0 fps; a frame: 824 draws (0 skipped), 104 dispatches, 142 passes, 144 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 37474 pipelines (7236569 ms), 0 programs (0 failed), 1239 textures (3519 MB)
[hos host] vk: host time in GPU resource waits: 0.50 ms a frame (fences 0.03, timelines 0.48, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 351.4 full barriers recorded, 184.6 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (RX_VK_PASS_BARRIER=game)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x8: 3.0; 0x9: 1.0; 0x11: 45.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 14.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 8.4 ms a frame (the longest submission 4.6 ms); 0.0 textures made a frame
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos Presentation Thread] vk: capture: a submit 0.14 ms (lock 0.00, scan 0.07 over 1278 ranges, 1763 MB), 208 pages written, 208 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (228 MB), 11 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: the recording streams: 48 of 2 MB, the most one held 346 KB; 0 made on demand (0 MB kept now); c1 stages: at most 114 a recording; 0 descriptor pools made when a recording's own ran out
[hos host] vk: sparse mirrors: 55, 787 MB with memory of 992 MB (204 MB spared); blocks given memory later 90, given back 1228; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 84 binds (11.3 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 7 taken from the spares, 63 made new; 15 kept as spares, 0 destroyed; 8 spares now, 156 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1289, 5276.2 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 264 dirty pages from their submits' captures, 66143 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): 1297 dispatches, 51466 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6734 runs, 914.2 MB compared; 124 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 23093 in captures, 4 only in the collection, 235 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 64 live (1092 MB), 0 retired still imported (0 MB), 8 retired given back (172 MB)
[hos host] vk: replay reads of guest memory as submitted: 1700000 (867408 pages from the captures)
[hos ModuleSystemWorker3] nvn: DrawMeshTasksIndirect: 140000 so far (x1 0x1184327040 x2 0x6001d40 x3 0x1184327000 x4 0xf33080)
[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 Presentation Thread] vi: 4987 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 1, 15.5-17.8 298, 17.8-22 2, 22-30 0, 30-40 0, over 40 0
[hos host] vk: replay reads of guest memory as submitted: 1720000 (877742 pages from the captures)
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1318 of 3000 waited for their write-back, 0.06 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 ModuleSystemWorker1] vk: invalidates since the last readbacks line: 4148 kept their virtual core (nothing there to wait for), 3020 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: mirror of pool 0x104bc00000 (4 MB) given back: idle, 6059 submissions since its sync
[hos host] vk: 16 mirrors given back (161 MB in all); 785 MB of mirrors now, 88 MB of them idle and kept for what the GPU wrote there alone
[hos sead::MainThread] nvn: EventSignal (CPU) #10000: event 0x1009b38004 mode 0 value 1 (lr 0x8348c78c)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07
[hos host] vk: write-backs left out (600 submits): 1288, 5274.3 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 316 dirty pages from their submits' captures, 66081 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): 1295 dispatches, 51436 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6619 runs, 898.1 MB compared; 143 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 22849 in captures, 5 only in the collection, 234 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: program picks: 1870187 kept (763538 by the last picks' textures, a uniform buffer bound since), 782 new, 1129030 with other textures (of 3000000)
[hos host] vk: replay reads of guest memory as submitted: 1740000 (888148 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.07 over 1278 ranges, 1763 MB), 208 pages written, 208 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.005; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (228 MB), 11 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: a frame: 51 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 6.92, present 0.25, draw 3.60, descriptors 1.42, texture lookups 0.07, uploads 0.00, pipelines 0.15, programs 0.71, commands 1.06, dispatch 0.82, memory sync 0.46, queue submit 0.51, slot wait 0.05
[hos host] vk: a frame: mirrors 70 (785 MB): 4.63 MB synced in 918.2 copies, 16.04 MB written back
[hos host] vk: 5040 presents, 60.0 fps; a frame: 822 draws (0 skipped), 104 dispatches, 141 passes, 144 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 37474 pipelines (7236569 ms), 0 programs (0 failed), 1240 textures (3524 MB)
[hos host] vk: host time in GPU resource waits: 0.57 ms a frame (fences 0.03, timelines 0.54, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 351.2 full barriers recorded, 184.8 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (RX_VK_PASS_BARRIER=game)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x8: 3.0; 0x9: 1.0; 0x11: 45.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 14.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 8.2 ms a frame (the longest submission 6.1 ms); 0.0 textures made a frame
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos host] vk: queues ordered by their events: 30000 submissions stopped at a wait, 29990 went on once it was reached, 8 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 2 (stopped 50.00 ms at most); waits met 89782, not met 0; captures at a CPU event's wait 1566, 0 of them at a wait met on arrival
[hos sead::MainThread] nvn: 550000 recordings made, 44336 live copies (12.8 MB), 319 texture copies; 4 command buffers' segments (4.2 MB)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #90000: event 0x10103c8048 mode 0 value 64847 (now 64828) (lr 0x806661f8)
[hos host] vk: sparse mirrors: 54, 785 MB with memory of 988 MB (202 MB spared); blocks given memory later 90, given back 1228; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 85 binds (11.4 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 7 taken from the spares, 63 made new; 16 kept as spares, 0 destroyed; 9 spares now, 160 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1288, 5274.3 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 216 dirty pages from their submits' captures, 66193 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): 1295 dispatches, 51425 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6632 runs, 898.2 MB compared; 174 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 22674 in captures, 2 only in the collection, 235 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 64 live (1092 MB), 0 retired still imported (0 MB), 8 retired given back (172 MB)
[hos host] vk: replay reads of guest memory as submitted: 1760000 (898538 pages from the captures)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #65000: event 0x10103c8048 mode 0 value 64999 (lr 0x80666598)
[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 0.08
[hos host] vk: replay reads of guest memory as submitted: 1780000 (908966 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 885615 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: write-backs left out (600 submits): 1288, 5274.3 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 241 dirty pages from their submits' captures, 66683 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): 1309 dispatches, 51829 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6627 runs, 898.2 MB compared; 198 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 22904 in captures, 3 only in the collection, 230 only by the replay itself
[hos host] vk:   read live, by pool:
[hos Presentation Thread] vk: capture: a submit 0.17 ms (lock 0.00, scan 0.08 over 1278 ranges, 1763 MB), 208 pages written, 208 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.007; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.003, bytes copied 0.000; 25 kept (220 MB), 11 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: replay reads of guest memory as submitted: 1800000 (919394 pages from the captures)
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1218 of 3000 waited for their write-back, 0.05 ms each; 0 looked at the stopped queues first, 1 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 ModuleSystemWorker3] vk: invalidates since the last readbacks line: 3851 kept their virtual core (nothing there to wait for), 3061 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: a frame: 51 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 7.70, present 0.29, draw 4.09, descriptors 1.66, texture lookups 0.08, uploads 0.00, pipelines 0.17, programs 0.77, commands 1.20, dispatch 0.90, memory sync 0.53, queue submit 0.58, slot wait 0.05
[hos host] vk: a frame: mirrors 71 (787 MB): 4.65 MB synced in 919.2 copies, 16.04 MB written back
[hos host] vk: 5160 presents, 60.0 fps; a frame: 821 draws (0 skipped), 104 dispatches, 141 passes, 144 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 37474 pipelines (7236569 ms), 0 programs (0 failed), 1240 textures (3524 MB)
[hos host] vk: host time in GPU resource waits: 0.32 ms a frame (fences 0.03, timelines 0.29, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 351.0 full barriers recorded, 184.7 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (RX_VK_PASS_BARRIER=game)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x8: 3.0; 0x9: 1.0; 0x11: 45.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 14.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 8.1 ms a frame (the longest submission 4.3 ms); 0.0 textures made a frame
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos PreActorModelCompose] vcores: 148076 takes, 20544 waits (mean 69 us, longest 8.1 ms, 0 over HOS_VCORE_LOG_MS), 10025 kicks, 9989 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: sparse mirrors: 55, 787 MB with memory of 992 MB (204 MB spared); blocks given memory later 90, given back 1228; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 86 binds (11.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 8 taken from the spares, 63 made new; 16 kept as spares, 0 destroyed; 8 spares now, 156 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1310, 5366.5 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 265 dirty pages from their submits' captures, 67112 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): 1304 dispatches, 52202 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6631 runs, 898.2 MB compared; 163 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 22669 in captures, 5 only in the collection, 232 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 64 live (1092 MB), 0 retired still imported (0 MB), 8 retired given back (172 MB)
[hos host] vk: replay reads of guest memory as submitted: 1820000 (929894 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 0.08
[hos host] vk: slow present: 19.0 ms:
[hos sead::MainThread] nvn: GPU waits in 5 s: 1 sync waits (6 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: present 5212 written to <home>\AppData\Local\<user>\data\jobs\jb_765b5a77c775d49d\runs\field-turn-and-stand-g2-01-pr\shots/field-turn-and-stand-pr-01-last-frame.png (the control bridge)
[hos host] vk: replay reads of guest memory as submitted: 1840000 (940088 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.15 ms (lock 0.00, scan 0.07 over 1278 ranges, 1763 MB), 208 pages written, 208 copied (0.8 MB)
[hos host] vk: a capture taken (ms): all 0.006; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.002, bytes copied 0.000; 25 kept (234 MB), 11 spare buffers, 0 texture pages copied (0 MB)
[hos host] vk: the recording streams: 48 of 2 MB, the most one held 346 KB; 0 made on demand (0 MB kept now); c1 stages: at most 114 a recording; 0 descriptor pools made when a recording's own ran out
[hos host] vk: write-backs left out (600 submits): 1289, 5276.2 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 363 dirty pages from their submits' captures, 66133 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): 1297 dispatches, 51501 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6734 runs, 914.1 MB compared; 192 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 22834 in captures, 149 only in the collection, 233 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: program picks: 1994847 kept (815025 by the last picks' textures, a uniform buffer bound since), 782 new, 1204370 with other textures (of 3200000)
[hos host] vk: replay reads of guest memory as submitted: 1860000 (950286 pages from the captures)
[hos sead::MainThread] nvn: CommandBufferWaitEvent #95000: event 0x10103c8048 mode 0 value 68647 (now 68628) (lr 0x806661f8)
[hos host] vk: a frame: 51 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 6.66, present 0.41, draw 3.47, descriptors 1.38, texture lookups 0.07, uploads 0.00, pipelines 0.15, programs 0.68, commands 1.01, dispatch 0.76, memory sync 0.46, queue submit 0.48, slot wait 0.05
[hos host] vk: a frame: mirrors 71 (787 MB): 4.64 MB synced in 919.6 copies, 16.04 MB written back
[hos host] vk: 5280 presents, 60.1 fps; a frame: 824 draws (0 skipped), 104 dispatches, 142 passes, 144 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 37474 pipelines (7236569 ms), 0 programs (0 failed), 1241 textures (3529 MB)
[hos host] vk: host time in GPU resource waits: 0.54 ms a frame (fences 0.03, timelines 0.51, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 351.4 full barriers recorded, 184.7 left out (nothing recorded since the one before), 0.0 pass starts ordered in their fragment stages only (RX_VK_PASS_BARRIER=game)
[hos host] vk: a frame: the game's barriers by their NVN flags: 0x1: 2.0; 0x8: 3.0; 0x9: 1.0; 0x11: 45.0; 0x12: 68.5; 0x13: 2.0; 0x21: 4.0; 0x28: 14.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 8.6 ms a frame (the longest submission 4.4 ms); 0.0 textures made a frame
[hos host] vk: a frame: 0.00 MB unswizzled on the CPU in 0.00 ms
[hos Presentation Thread] vi: 5288 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 47, 15.5-17.8 219, 17.8-22 33, 22-30 2, 30-40 0, over 40 0
[hos host] vk: sparse mirrors: 55, 787 MB with memory of 992 MB (204 MB spared); blocks given memory later 90, given back 1228; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 0; 86 binds (11.5 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 8 taken from the spares, 63 made new; 16 kept as spares, 0 destroyed; 8 spares now, 156 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 1288, 5274.3 MB; 31 bindings followed, 46 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 0 so far
[hos host] vk: mirror syncs (600 submits): 347 dirty pages from their submits' captures, 66052 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): 1295 dispatches, 51379 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6639 runs, 898.2 MB compared; 129 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 23053 in captures, 2 only in the collection, 236 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 64 live (1092 MB), 0 retired still imported (0 MB), 8 retired given back (172 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: a submit's memory work (ms): maps 0.00, capture taken 0.01, writes collected 0.07
[hos host] vk: replay reads of guest memory as submitted: 1880000 (960502 pages from the captures)
[hos ModuleSystemWorker3] vk: readbacks of GPU results: 1336 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 ModuleSystemWorker3] vk: invalidates since the last readbacks line: 4143 kept their virtual core (nothing there to wait for), 3025 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: 1 textures retired (1240 left, 3524 MB): budget 0, drawn over 1, dead targets 0