Beanstalk Longshot test farm
Menu

Report rp_b31287ccf80ab4b1 pass

#1488 perf(vk): let go of the shader modules of ready programs after the start on Windows too, on request (#1474)

linux on Alberania/agent-af6cf3,

This report disagrees with its peers.

module-release-tour run_frametime_max_ms: +21.4% against peers' -60.3%
module-release-tour#1 market.hitch_count: -50.0% against peers' +100.0%
module-release-tour#1 market.stutter_count: -33.3% against peers' +100.0%
module-release-tour#1 run_frametime_max_ms: +2.3% against peers' -70.8%

Machine

Commit
ce69221b74cfa84b683fd147eefae9c4645c32a4
Baseline
068fb8ebe232d455ef48e817702b03ccc62d8520
App
nightly-191-38dff5fc, attempt 2
Game build
e0f3ccfa
os
linux
cpu
AMD Ryzen 7 5800X3D 8-Core Processor
gpu
AMD Radeon RX 7900 XTX (RADV NAVI31)
tags
linuxamdradvdgpu
driver
26.2.7
ram_gb
64
vram_mb
24,560
gpu_vendor
amd
os_version
Bazzite (kernel 7.3.4)

Builds

main (baseline)

0 s warm

from 068fb8ebe232

the base was built already

PR

23.4 s warm

Program cache: warm

Scenarios

module-release-tour RX_VK_DROP_MODULES=0

ok
Metric (median)mainPRΔ
exit_code00±0%
field.cpu_s40.8840.56−0.8%
load_time_s4.634.87+5.2%
market.cpu_s70.4273.7+4.7%
village.cpu_s59.2960.06+1.3%
load_time_max_s5.485.51+0.5%
run_hitch_count4650+8.7%
field.hitch_count00±0%
field.rss_peak_mb12,153.112,270.4+1%
log_forbidden_hits00±0%
market.hitch_count01—
market.rss_peak_mb13,551.113,816.7+2%
field.stutter_count00±0%
village.hitch_count22±0%
village.rss_peak_mb13,362.313,519.5+1.2%
log_expected_missing00±0%
market.stutter_count02—
run_frametime_max_ms274.07332.75+21.4%
village.stutter_count22±0%
field.frametime_max_ms28.4424.58−13.6%
field.frametime_p50_ms16.6516.59−0.4%
field.frametime_p99_ms19.2118.68−2.8%
value_modules_released00±0%
value_private_after_mb00±0%
market.frametime_max_ms30.0673.96+146%
market.frametime_p50_ms16.9316.37−3.3%
market.frametime_p99_ms20.1820.4+1.1%
value_private_before_mb00±0%
village.frametime_max_ms72.4975.81+4.6%
village.frametime_p50_ms16.4216.42±0%
village.frametime_p99_ms21.7822.1+1.5%
field.frametime_1pct_low_ms20.8919.74−5.5%
market.frametime_1pct_low_ms22.6929.56+30.3%
village.frametime_1pct_low_ms30.9132.06+3.7%
screenshot_diff—0—

Pictures against main: field diff 0 (main against itself ), market diff 0 (main against itself ), village diff 0 (main against itself )

Runs (2)
Metricmain 0pr 1
hangnono
crashednono
exit_code00
field.cpu_s40.8840.56
load_time_s4.634.87
market.cpu_s70.4273.7
village.cpu_s59.2960.06
load_time_max_s5.485.51
run_hitch_count4650
field.hitch_count00
field.rss_peak_mb12,153.112,270.4
log_forbidden_hits00
market.hitch_count01
market.rss_peak_mb13,551.113,816.7
field.stutter_count00
village.hitch_count22
village.rss_peak_mb13,362.313,519.5
log_expected_missing00
market.stutter_count02
run_frametime_max_ms274.07332.75
village.stutter_count22
field.frametime_max_ms28.4424.58
field.frametime_p50_ms16.6516.59
field.frametime_p99_ms19.2118.68
value_modules_released00
value_private_after_mb00
market.frametime_max_ms30.0673.96
market.frametime_p50_ms16.9316.37
market.frametime_p99_ms20.1820.4
value_private_before_mb00
village.frametime_max_ms72.4975.81
village.frametime_p50_ms16.4216.42
village.frametime_p99_ms21.7822.1
field.frametime_1pct_low_ms20.8919.74
market.frametime_1pct_low_ms22.6929.56
village.frametime_1pct_low_ms30.9132.06
screenshot_diff—0
loadsHyruleField_Child 3.86 sMarket_Child 5.48 sKakarikoVillage_Child 4.56 sHyruleField_Child 4.2 sMarket_Child 5.51 sKakarikoVillage_Child 4.9 s

module-release-tour RX_VK_DROP_MODULES=1

ok
Metric (median)mainPRΔ
exit_code00±0%
field.cpu_s40.2640.23−0.1%
load_time_s4.84.78−0.4%
market.cpu_s70.4170.91+0.7%
village.cpu_s59.659.01−1%
load_time_max_s5.355.25−1.9%
run_hitch_count4451+15.9%
field.hitch_count00±0%
field.rss_peak_mb7,947.77,867.6−1%
log_forbidden_hits00±0%
market.hitch_count21−50%
market.rss_peak_mb9,372.19,398.3+0.3%
field.stutter_count00±0%
village.hitch_count11±0%
village.rss_peak_mb9,145.39,109.2−0.4%
log_expected_missing00±0%
market.stutter_count32−33.3%
run_frametime_max_ms365.46373.95+2.3%
village.stutter_count12+100%
field.frametime_max_ms30.1929.12−3.5%
field.frametime_p50_ms16.6516.61−0.3%
field.frametime_p99_ms18.0218.28+1.4%
value_modules_released62,38462,384±0%
value_private_after_mb01,961—
market.frametime_max_ms68.6164.92−5.4%
market.frametime_p50_ms16.916.68−1.3%
market.frametime_p99_ms19.9819.77−1.1%
value_private_before_mb015,505—
village.frametime_max_ms79.0475.42−4.6%
village.frametime_p50_ms16.3816.48+0.6%
village.frametime_p99_ms21.5921.24−1.6%
field.frametime_1pct_low_ms20.2319.98−1.2%
market.frametime_1pct_low_ms29.7126.68−10.2%
village.frametime_1pct_low_ms29.128.24−2.9%
screenshot_diff—0—

Pictures against main: field diff 0 (main against itself ), market diff 0 (main against itself ), village diff 0 (main against itself )

Runs (2)
Metricmain 0pr 1
hangnono
crashednono
exit_code00
field.cpu_s40.2640.23
load_time_s4.84.78
market.cpu_s70.4170.91
village.cpu_s59.659.01
load_time_max_s5.355.25
run_hitch_count4451
field.hitch_count00
field.rss_peak_mb7,947.77,867.6
log_forbidden_hits00
market.hitch_count21
market.rss_peak_mb9,372.19,398.3
field.stutter_count00
village.hitch_count11
village.rss_peak_mb9,145.39,109.2
log_expected_missing00
market.stutter_count32
run_frametime_max_ms365.46373.95
village.stutter_count12
field.frametime_max_ms30.1929.12
field.frametime_p50_ms16.6516.61
field.frametime_p99_ms18.0218.28
value_modules_released62,38462,384
value_private_after_mb01,961
market.frametime_max_ms68.6164.92
market.frametime_p50_ms16.916.68
market.frametime_p99_ms19.9819.77
value_private_before_mb015,505
village.frametime_max_ms79.0475.42
village.frametime_p50_ms16.3816.48
village.frametime_p99_ms21.5921.24
field.frametime_1pct_low_ms20.2319.98
market.frametime_1pct_low_ms29.7126.68
village.frametime_1pct_low_ms29.128.24
screenshot_diff—0
loadsHyruleField_Child 4.15 sMarket_Child 5.35 sKakarikoVillage_Child 4.89 sHyruleField_Child 4.17 sMarket_Child 5.25 sKakarikoVillage_Child 4.92 s

Rules

ScenarioCellRuleLimitmainPRΔ %Status
module-release-tourRX_VK_DROP_MODULES=0crashFalse00—pass
module-release-tourRX_VK_DROP_MODULES=0screenshot.max_diff0.15—0—pass
module-release-tourRX_VK_DROP_MODULES=0field.rss_peak_mb.max_delta_pct1512,153.112,270.4+1%pass
module-release-tourRX_VK_DROP_MODULES=0market.rss_peak_mb.max_delta_pct1513,551.113,816.7+2%pass
module-release-tourRX_VK_DROP_MODULES=0village.rss_peak_mb.max_delta_pct1513,362.313,519.5+1.2%pass
module-release-tourRX_VK_DROP_MODULES=0load_time_max_s.max_delta_pct305.485.51+0.5%pass
module-release-tourRX_VK_DROP_MODULES=0field.frametime_p99_ms.max_delta_pct1519.2118.68−2.8%pass
module-release-tourRX_VK_DROP_MODULES=0market.frametime_p99_ms.max_delta_pct1520.1820.4+1.1%pass
module-release-tourRX_VK_DROP_MODULES=0village.frametime_p99_ms.max_delta_pct1521.7822.1+1.5%pass
module-release-tourRX_VK_DROP_MODULES=0log_forbidden_hits.max000—pass
module-release-tourRX_VK_DROP_MODULES=1crashFalse00—pass
module-release-tourRX_VK_DROP_MODULES=1screenshot.max_diff0.15—0—pass
module-release-tourRX_VK_DROP_MODULES=1field.rss_peak_mb.max_delta_pct157,947.77,867.6−1%pass
module-release-tourRX_VK_DROP_MODULES=1market.rss_peak_mb.max_delta_pct159,372.19,398.3+0.3%pass
module-release-tourRX_VK_DROP_MODULES=1village.rss_peak_mb.max_delta_pct159,145.39,109.2−0.4%pass
module-release-tourRX_VK_DROP_MODULES=1load_time_max_s.max_delta_pct305.355.25−1.9%pass
module-release-tourRX_VK_DROP_MODULES=1field.frametime_p99_ms.max_delta_pct1518.0218.28+1.4%pass
module-release-tourRX_VK_DROP_MODULES=1market.frametime_p99_ms.max_delta_pct1519.9819.77−1.1%pass
module-release-tourRX_VK_DROP_MODULES=1village.frametime_p99_ms.max_delta_pct1521.5921.24−1.6%pass
module-release-tourRX_VK_DROP_MODULES=1log_forbidden_hits.max000—pass
module-release-tourRX_VK_DROP_MODULES=1field.frametime_p99_ms: RX_VK_DROP_MODULES=1 vs RX_VK_DROP_MODULES=0 (PR)1518.6818.28−2.1%passthe PR's own cells: RX_VK_DROP_MODULES=1 against RX_VK_DROP_MODULES=0 (RX_VK_DROP_MODULES=0)
module-release-tourRX_VK_DROP_MODULES=1market.frametime_p99_ms: RX_VK_DROP_MODULES=1 vs RX_VK_DROP_MODULES=0 (PR)1520.419.77−3.1%passthe PR's own cells: RX_VK_DROP_MODULES=1 against RX_VK_DROP_MODULES=0 (RX_VK_DROP_MODULES=0)
module-release-tourRX_VK_DROP_MODULES=1village.frametime_p99_ms: RX_VK_DROP_MODULES=1 vs RX_VK_DROP_MODULES=0 (PR)1522.121.24−3.9%passthe PR's own cells: RX_VK_DROP_MODULES=1 against RX_VK_DROP_MODULES=0 (RX_VK_DROP_MODULES=0)
module-release-tourRX_VK_DROP_MODULES=1run_hitch_count: RX_VK_DROP_MODULES=1 vs RX_VK_DROP_MODULES=0 (PR)505051+2%passthe PR's own cells: RX_VK_DROP_MODULES=1 against RX_VK_DROP_MODULES=0 (RX_VK_DROP_MODULES=0)

Log (PR side, end)

[hos host] vk:   read live, by pool:
[hos host] vk: imports: 150 live (2328 MB), 0 retired still imported (0 MB), 53 retired given back (1408 MB)
[hos host] vk: replay reads of guest memory as submitted: 3680000 (2076831 pages from the captures)
[hos ModuleSystemWorker4] nvn: CommandBufferSignalEvent #80000: event 0x100d6c801c mode 0 value 79997 (lr 0x802911b8)
[hos host] vk: replay reads of guest memory as submitted: 3700000 (2087749 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 640, 2795.9 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2975 dirty pages from their submits' captures, 41285 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): 807 dispatches, 32818 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6395 runs, 595.9 MB compared; 58 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 15039 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos sead::MainThread] vk: readbacks of GPU results: 1337 of 3000 waited for their write-back, 0.47 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: 4212 kept their virtual core (nothing there to wait for), 2956 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: 3720000 (2098085 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.03
[hos host] vk: memory: working set 9103 MB (peak 16761), private 2532 MB, staging 256 MB in 16 arenas, 36 made so far (0 ms)
[hos host] vk: memory heaps: heap 0 (system): 20 blocks 968 MB, 39 allocations 902 MB; used 3320 of 31967 MB; heap 1 (device): 338 blocks 11282 MB, 20653 allocations 10556 MB; used 12004 of 22637 MB;
[hos host] vk: image blocks: 1 blocks 32 MB, 5 images 16 MB; 2 free ranges, the largest 15 MB
[hos host] vk: image blocks (render targets): 19 blocks 608 MB, 126 images 344 MB; 80 free ranges, the largest 31 MB
[hos host] vk: image blocks (textures, compacted): 217 blocks 6944 MB, 4432 images 6857 MB; 993 free ranges, the largest 19 MB
[hos host] vk: render target images: 573 reused from the pool, 572 made new, 440 destroyed from it; it holds 0 (0 MB)
[hos host] vk: VRAM: 12004 MB used of 22637 MB allowed (726 MB of it free in the blocks); textures 8749 MB (budget 11000 MB), 4574 textures
[hos host] vk: budget: VRAM 11282 of 22637 MB (12004 MB by the driver, 721 MB unaccounted): textures 8749, mirrors 986, blocks free 726, shadows 304, staging 256, other buffers 227, other images 31; RAM 3265 MB (the process's private memory 2532 MB): imports 2328, capture spares 742, captures 150, capture table 35, other buffers 10, others 0
[hos host] vk: textures: decoded ASTC 5782 MB, rendered 1101 MB (106 drawn in one frame, kept: 51680 KB), other 1102 MB; used in the last 8 presents 2379 MB
[hos host] vk: decoded ASTC used again: 155 textures (397 MB) copied from ones the game finalized, not decoded; 155 of those retired, 715 listed; 1190 MB of blocks hashed in 114 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 1 presents ago
[hos host] vk:   214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, 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:   64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk:   64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0xa, last used 4577 presents ago
[hos host] vk:   54 MB: 0x10005552400 3072x3072x1 x3 levels 1 NVN 0x32 rendered, last used 2671 presents ago
[hos host] vk:   54 MB: 0x10005191c00 3072x3072x1 x3 levels 1 NVN 0x32 rendered, last used 2688 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 13 (2113 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 260x260x1 NVN 0xd: 12 (2111 KB);
[hos Presentation Thread] vk: presents after the CPU's signal (600): 521 waited for it, 2.90 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 80.7 ms; VRAM 11.7 of 22.1 GB
[hos host] vk: a frame: 191 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 73 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 8.10, present 0.16, draw 4.79, descriptors 1.95, texture lookups 0.05, uploads 0.00, pipelines 0.20, programs 0.75, commands 1.53, dispatch 0.57, memory sync 0.36, queue submit 1.00, slot wait 0.06
[hos host] vk: a frame: mirrors 220 (986 MB): 5.39 MB synced in 1022.7 copies, 18.62 MB written back
[hos host] vk: 6600 presents, 59.2 fps; a frame: 2155 draws (0 skipped), 168 dispatches, 187 passes, 146 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 37379 pipelines (989 ms), 0 programs (10 failed), 4574 textures (8749 MB)
[hos host] vk: host time in GPU resource waits: 6.65 ms a frame (fences 0.00, timelines 6.65, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 665.9 full barriers recorded, 267.6 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: 42.0; 0x12: 73.5; 0x13: 2.0; 0x21: 4.0; 0x28: 14.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 11.9 ms a frame (the longest submission 4.7 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, 16.81 ms a frame): its jobs 8.75 ms a frame (submits 1.04, presents 0.16, others 7.55), idle 8.07; the game waited for room at 1 presents, 0.36 ms each (0.00 a frame; 1.87 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); 585 captures at a CPU signal; the submits' waits read in 0.052 ms a frame; readbacks: 2999 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 7.94 ms after its hand-over (24.55 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 1.81 + 0.25 (the window's copy is in the idle), after it 2.56 + 0.38; after a frame's last recording that reads or writes the game's memory come 1.02 recordings (0.02 ms of GPU time), after its last that reads it 1.02 (0.02 ms)
[hos host] vk: frame ahead (600 presents): 11.77 parts a frame recorded in the signal's job, 83.4% of them took nothing that came with the signal; of the others, 1.96 a frame took pages of the signal's capture into a mirror (91.3 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 2 of 584 frames, its first three in 2
[hos host] vk: frame ahead, what came with the signal: 585 signals with a capture, each of 118.1 pages in 9.9 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: 0.57+0.03, 0.05+0.02, 0.10+0.04, 0.20+0.11, 0.23+0.03, 0.37+0.02, 0.03+0.01, 0.39+0.04, 0.21+0.01, 0.20+0.02, 0.21+0.03, 0.00+0.01
[hos host] vk: replay reads of guest memory as submitted: 3740000 (2108403 pages from the captures)
[hos host] vk: sparse mirrors: 87, 986 MB with memory of 1540 MB (553 MB spared); blocks given memory later 8124, given back 14532; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 9054; 622 binds (340.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 86 taken from the spares, 134 made new; 133 kept as spares, 0 destroyed; 47 spares now, 1396 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 640, 2795.9 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2539 dirty pages from their submits' captures, 41443 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): 819 dispatches, 32773 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6405 runs, 596.5 MB compared; 58 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 14954 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 150 live (2328 MB), 0 retired still imported (0 MB), 53 retired given back (1408 MB)
[hos nn.npln.Shared] vcores: 86086 takes, 13108 waits (mean 45 us, longest 3.3 ms, 0 over HOS_VCORE_LOG_MS), 5430 kicks, 5634 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: program picks: 5939209 kept (2296882 by the last picks' textures, a uniform buffer bound since), 23316 new, 2037474 with other textures (of 8000000)
[hos host] vk: late pages' fix (600 recordings with one): of 613 captures with late pages 612 fixed (164.1 pages, 165.0 fixes a capture), waited as before: 0 no pool, 1 no stamps or shadow yet, 0 stamps too new, 0 whole copies, 0 in place or unknown, 0 a page's earlier fix not recorded yet, 0 other; 570 read on the CPU first, 0 with no table, 0 pages whose shadow changed first; stamps 102 blocks (6528 KB); of the reads on the CPU, 0 waited for an earlier part's fix, 0 copied the page again instead (that part not waited for); the paths: reads on the CPU with the fix done 391, of an earlier recording 179, not recorded yet 0, of this part 0; pages fixed from stamps 99208, from snapshots 1197; the shadows: 320678 pages synced from captures, 317985 made known by it, 2693 left unknown because the replay or the collection wrote them since
[hos Presentation Thread] vk: capture: a submit 0.11 ms (lock 0.00, scan 0.04 over 669 ranges, 3867 MB), 245 pages written, 245 copied (1.0 MB)
[hos host] vk: a capture taken (ms): all 0.025; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.016, bytes copied 0.000; 24 kept (156 MB), 12 spare buffers, 7799 texture pages copied (30 MB)
[hos host] vk: replay reads of guest memory as submitted: 3760000 (2118744 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 34512 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: readback pages refreshed from guest memory: 1656382 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: recorder: 8122094 packets run in 10.0 s (busy 1.26 s, spun idle 3.70 s, slept 4893 times, 1995 of them under 200 us), 0 calls made directly (no room); the replay waited for it 30602 times, 68.27 ms in all, the longest 0.18 ms; vkEndCommandBuffer 11201/31.28 ms (max 0.18), vkResetDescriptorPool 11162/17.79 ms (max 0.02), replay_mu unlock 7745/17.77 ms (max 0.12), vmaInvalidateAllocation 352/1.05 ms (max 0.17), vkQueueSubmit2 142/0.38 ms (max 0.15)
[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: the recording streams: 48 of 2 MB, the most one held 347 KB; 0 made on demand (0 MB kept now); c1 stages: at most 168 a recording; 0 descriptor pools made when a recording's own ran out
[hos host] vk: write-backs left out (600 submits): 640, 2795.9 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 3132 dirty pages from their submits' captures, 41056 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): 807 dispatches, 32764 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6385 runs, 595.9 MB compared; 59 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 15066 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 3780000 (2129029 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 3800000 (2139332 pages from the captures)
[hos Presentation Thread] nvn: QueueSubmitCommands #40000: queue 0x101004133020 (its #6674), 1 handles (lr 0x80590358)
[hos host] vk: sparse mirrors: 87, 986 MB with memory of 1540 MB (553 MB spared); blocks given memory later 8124, given back 14532; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 9054; 622 binds (340.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 86 taken from the spares, 134 made new; 133 kept as spares, 0 destroyed; 47 spares now, 1396 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 639, 2794.1 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2948 dirty pages from their submits' captures, 41180 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): 804 dispatches, 32763 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6390 runs, 596.4 MB compared; 56 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 15035 in captures, 0 only in the collection, 31 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 150 live (2328 MB), 0 retired still imported (0 MB), 53 retired given back (1408 MB)
[hos host] vk: replay reads of guest memory as submitted: 3820000 (2149674 pages from the captures)
[hos sead::MainThread] nvn: GPU waits in 5 s: 264 sync waits (715 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos ModuleSystemWorker3] nvn: CommandBufferWaitEvent #115000: event 0x100d6c8018 mode 0 value 82333 (now 82314) (lr 0x80661124)
[hos host] vk: replay reads of guest memory as submitted: 3840000 (2159991 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.03
[hos host] vk: program picks: 6087307 kept (2356877 by the last picks' textures, a uniform buffer bound since), 23316 new, 2089376 with other textures (of 8200000)
[hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.04 over 667 ranges, 3859 MB), 222 pages written, 222 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.016; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.006, bytes copied 0.000; 24 kept (146 MB), 12 spare buffers, 7801 texture pages copied (30 MB)
[hos host] vk: write-backs left out (600 submits): 640, 2795.9 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2952 dirty pages from their submits' captures, 41131 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): 802 dispatches, 32764 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6385 runs, 597.6 MB compared; 62 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 15058 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 3860000 (2170341 pages from the captures)
[hos host] vk: a frame: 191 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 73 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 8.14, present 0.15, draw 4.85, descriptors 1.98, texture lookups 0.06, uploads 0.00, pipelines 0.20, programs 0.76, commands 1.55, dispatch 0.58, memory sync 0.32, queue submit 1.00, slot wait 0.07
[hos host] vk: a frame: mirrors 220 (986 MB): 5.39 MB synced in 1021.1 copies, 18.65 MB written back
[hos host] vk: 6720 presents, 60.0 fps; a frame: 2151 draws (0 skipped), 168 dispatches, 187 passes, 146 barriers, 6 submits, 0.1 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 37379 pipelines (989 ms), 0 programs (10 failed), 4580 textures (8750 MB)
[hos host] vk: host time in GPU resource waits: 6.17 ms a frame (fences 0.00, timelines 6.17, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 666.3 full barriers recorded, 267.7 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: 42.0; 0x12: 73.5; 0x13: 2.0; 0x21: 4.0; 0x28: 14.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 11.8 ms a frame (the longest submission 4.6 ms); 0.1 textures made a frame
[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: 3880000 (2180622 pages from the captures)
[hos host] vk: sparse mirrors: 87, 986 MB with memory of 1540 MB (553 MB spared); blocks given memory later 8124, given back 14532; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 9054; 622 binds (340.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 86 taken from the spares, 134 made new; 133 kept as spares, 0 destroyed; 47 spares now, 1396 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 621, 2710.2 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 3098 dirty pages from their submits' captures, 40467 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): 803 dispatches, 32211 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6243 runs, 579.2 MB compared; 53 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 15050 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 150 live (2328 MB), 0 retired still imported (0 MB), 53 retired given back (1408 MB)
[hos ModuleSystemWorker4] vk: readbacks of GPU results: 1350 of 3000 waited for their write-back, 0.47 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: 4545 kept their virtual core (nothing there to wait for), 3135 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: 3900000 (2190934 pages from the captures)
[hos Presentation Thread] vi: 6766 frames presented (59.7 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 79, 15.5-17.8 170, 17.8-22 46, 22-30 3, 30-40 1, over 40 0
[hos host] vk: replay reads of guest memory as submitted: 3920000 (2201286 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 640, 2795.9 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2838 dirty pages from their submits' captures, 41178 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): 806 dispatches, 32740 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6404 runs, 595.9 MB compared; 68 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 14953 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 3940000 (2211582 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 1726869 pages so far (0 held: may be written ahead in their submission)
[hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.04 over 667 ranges, 3859 MB), 226 pages written, 226 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.019; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.009, bytes copied 0.000; 24 kept (140 MB), 12 spare buffers, 3975 texture pages copied (15 MB)
[hos host] vk: program picks: 6235563 kept (2417075 by the last picks' textures, a uniform buffer bound since), 23316 new, 2141120 with other textures (of 8400000)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.03
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 15442 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: the recording streams: 48 of 2 MB, the most one held 347 KB; 0 made on demand (0 MB kept now); c1 stages: at most 168 a recording; 0 descriptor pools made when a recording's own ran out
[hos host] vk: sparse mirrors: 87, 986 MB with memory of 1540 MB (553 MB spared); blocks given memory later 8124, given back 14532; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 9054; 622 binds (340.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 86 taken from the spares, 134 made new; 133 kept as spares, 0 destroyed; 47 spares now, 1396 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 640, 2795.9 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2680 dirty pages from their submits' captures, 41302 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): 804 dispatches, 32714 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6403 runs, 597.0 MB compared; 55 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 14465 in captures, 1 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 150 live (2328 MB), 0 retired still imported (0 MB), 53 retired given back (1408 MB)
[hos host] vk: queues ordered by their events: 42000 submissions stopped at a wait, 41818 went on once it was reached, 180 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 18 (stopped 132.41 ms at most); waits met 117602, not met 0; captures at a CPU event's wait 4267, 0 of them at a wait met on arrival
[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: 3960000 (2221959 pages from the captures)
[hos ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 200000 so far (x1 0x1179846000 x2 0 x3 0x1179846000 x4 0x654040)
[hos host] vk: replay reads of guest memory as submitted: 3980000 (2232236 pages from the captures)
[hos ModuleSystemWorker4] nvn: CommandBufferSignalEvent #85000: event 0x100d6c8030 mode 0 value 84999 (lr 0x806612a8)
[hos host] vk: write-backs left out (600 submits): 640, 2795.9 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2297 dirty pages from their submits' captures, 40865 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): 792 dispatches, 32288 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6394 runs, 595.9 MB compared; 62 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 14951 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 4000000 (2242550 pages from the captures)
[hos host] vk: render targets drawn over from another start, let go: 905 so far (4101 MB), 767 of them (3480 MB) between the passes
[hos host] vk: a frame: 192 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 73 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 8.17, present 0.15, draw 4.90, descriptors 2.05, texture lookups 0.04, uploads 0.00, pipelines 0.20, programs 0.75, commands 1.54, dispatch 0.59, memory sync 0.31, queue submit 1.02, slot wait 0.06
[hos host] vk: a frame: mirrors 220 (986 MB): 5.37 MB synced in 1020.1 copies, 18.64 MB written back
[hos host] vk: 6840 presents, 60.0 fps; a frame: 2156 draws (0 skipped), 168 dispatches, 187 passes, 146 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.1 ms (recording slots only); 37379 pipelines (989 ms), 0 programs (10 failed), 4580 textures (8750 MB)
[hos host] vk: host time in GPU resource waits: 5.86 ms a frame (fences 0.00, timelines 5.86, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 665.8 full barriers recorded, 267.6 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: 42.0; 0x12: 73.5; 0x13: 2.0; 0x21: 4.0; 0x28: 14.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 11.8 ms a frame (the longest submission 4.5 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: replay reads of guest memory as submitted: 4020000 (2252900 pages from the captures)
[hos ModuleSystemWorker2] nvn: 950000 recordings made, 57212 live copies (20.3 MB), 656 texture copies; 6 command buffers' segments (4.3 MB)
[hos host] vk: sparse mirrors: 87, 986 MB with memory of 1540 MB (553 MB spared); blocks given memory later 8124, given back 14532; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 9054; 622 binds (340.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 86 taken from the spares, 134 made new; 133 kept as spares, 0 destroyed; 47 spares now, 1396 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 640, 2795.9 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2610 dirty pages from their submits' captures, 42369 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): 827 dispatches, 33256 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6394 runs, 595.3 MB compared; 63 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 15542 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 150 live (2328 MB), 0 retired still imported (0 MB), 53 retired given back (1408 MB)
[hos host] vk: replay reads of guest memory as submitted: 4040000 (2263706 pages from the captures)
[hos sead::MainThread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.04 over 667 ranges, 3859 MB), 216 pages written, 216 copied (0.8 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 0.004, bytes copied 0.000; 25 kept (134 MB), 11 spare buffers, 4140 texture pages copied (16 MB)
[hos host] vk: program picks: 6383579 kept (2477040 by the last picks' textures, a uniform buffer bound since), 23316 new, 2193104 with other textures (of 8600000)
[hos host] vk: replay reads of guest memory as submitted: 4060000 (2274046 pages from the captures)
[hos ModuleSystemWorker5] nvn: CommandBufferWaitEvent #120000: event 0x100d6c8018 mode 0 value 86133 (now 86114) (lr 0x80661124)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.03
[hos host] vk: write-backs left out (600 submits): 659, 2881.5 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2167 dirty pages from their submits' captures, 42318 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): 813 dispatches, 33338 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6399 runs, 596.5 MB compared; 66 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 14900 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos sead::MainThread] vk: readbacks of GPU results: 1371 of 3000 waited for their write-back, 0.44 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: 4279 kept their virtual core (nothing there to wait for), 2889 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos rmbl::VibrationThread] hid: 20480 vibration values, strongest 0.00 since the last count
[hos host] vk: replay reads of guest memory as submitted: 4080000 (2284423 pages from the captures)
[hos ModuleSystemWorker5] vcores: 87590 takes, 13073 waits (mean 46 us, longest 4.3 ms, 0 over HOS_VCORE_LOG_MS), 5548 kicks, 5806 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 4100000 (2294815 pages from the captures)
[hos host] vk: sparse mirrors: 87, 986 MB with memory of 1540 MB (553 MB spared); blocks given memory later 8124, given back 14532; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 9054; 622 binds (340.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 86 taken from the spares, 134 made new; 133 kept as spares, 0 destroyed; 47 spares now, 1396 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 621, 2710.2 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2581 dirty pages from their submits' captures, 40843 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): 808 dispatches, 32216 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6411 runs, 596.7 MB compared; 56 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 14994 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 150 live (2328 MB), 0 retired still imported (0 MB), 53 retired given back (1408 MB)
[hos host] vk: replay reads of guest memory as submitted: 4120000 (2305167 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 1796476 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 6397 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: 191 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 73 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 8.13, present 0.16, draw 4.77, descriptors 1.93, texture lookups 0.05, uploads 0.00, pipelines 0.21, programs 0.74, commands 1.53, dispatch 0.59, memory sync 0.29, queue submit 1.02, slot wait 0.06
[hos host] vk: a frame: mirrors 220 (986 MB): 5.38 MB synced in 1021.9 copies, 18.67 MB written back
[hos host] vk: 6960 presents, 59.8 fps; a frame: 2156 draws (0 skipped), 168 dispatches, 187 passes, 146 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 37379 pipelines (989 ms), 0 programs (10 failed), 4580 textures (8750 MB)
[hos host] vk: host time in GPU resource waits: 5.62 ms a frame (fences 0.00, timelines 5.62, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 665.4 full barriers recorded, 267.5 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: 42.0; 0x12: 73.5; 0x13: 2.0; 0x21: 4.0; 0x28: 14.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 11.8 ms a frame (the longest submission 4.5 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 0.10 ms (lock 0.00, scan 0.04 over 667 ranges, 3853 MB), 215 pages written, 215 copied (0.8 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 0.004, bytes copied 0.000; 24 kept (128 MB), 12 spare buffers, 4260 texture pages copied (16 MB)
[hos host] vk: replay reads of guest memory as submitted: 4140000 (2315506 pages from the captures)
[hos host] vk: the recording streams: 48 of 2 MB, the most one held 347 KB; 0 made on demand (0 MB kept now); c1 stages: at most 168 a recording; 0 descriptor pools made when a recording's own ran out
[hos host] vk: write-backs left out (600 submits): 640, 2795.9 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2874 dirty pages from their submits' captures, 40548 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): 791 dispatches, 32334 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6443 runs, 600.7 MB compared; 61 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 14538 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: program picks: 6531420 kept (2536964 by the last picks' textures, a uniform buffer bound since), 23316 new, 2245263 with other textures (of 8800000)
[hos host] vk: replay reads of guest memory as submitted: 4160000 (2325885 pages from the captures)
[hos sead::MainThread] nvn: GPU waits in 5 s: 246 sync waits (610 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] vk: sparse mirrors: 87, 986 MB with memory of 1540 MB (553 MB spared); blocks given memory later 8124, given back 14532; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 9054; 622 binds (340.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 86 taken from the spares, 134 made new; 133 kept as spares, 0 destroyed; 47 spares now, 1396 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 640, 2795.9 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2758 dirty pages from their submits' captures, 41335 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): 805 dispatches, 32774 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6408 runs, 597.1 MB compared; 60 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 14946 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 150 live (2328 MB), 0 retired still imported (0 MB), 53 retired given back (1408 MB)
[hos host] vk: replay reads of guest memory as submitted: 4180000 (2336301 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.03
[hos host] vk: replay reads of guest memory as submitted: 4200000 (2346630 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 640, 2795.9 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2971 dirty pages from their submits' captures, 41204 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): 808 dispatches, 32783 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6408 runs, 598.6 MB compared; 57 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 15056 in captures, 0 only in the collection, 31 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 4220000 (2357030 pages from the captures)
[hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.04 over 667 ranges, 3853 MB), 234 pages written, 234 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.020; GPU wait for late pages 0.000 (0 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 24 kept (114 MB), 12 spare buffers, 6028 texture pages copied (23 MB)
[hos host] vk: replay reads of guest memory as submitted: 4240000 (2367462 pages from the captures)
[hos ModuleSystemWorker1] vk: readbacks of GPU results: 1382 of 3000 waited for their write-back, 0.46 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: 4641 kept their virtual core (nothing there to wait for), 3039 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: program picks: 6679674 kept (2596739 by the last picks' textures, a uniform buffer bound since), 23316 new, 2297009 with other textures (of 9000000)
[hos host] vk: sparse mirrors: 87, 986 MB with memory of 1540 MB (553 MB spared); blocks given memory later 8124, given back 14532; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 9054; 622 binds (340.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 86 taken from the spares, 134 made new; 133 kept as spares, 0 destroyed; 47 spares now, 1396 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 639, 2794.1 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2880 dirty pages from their submits' captures, 41291 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): 802 dispatches, 32840 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6410 runs, 597.8 MB compared; 56 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 15094 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 150 live (2328 MB), 0 retired still imported (0 MB), 53 retired given back (1408 MB)
[hos Presentation Thread] vi: 7067 frames presented (60.0 fps over the last 5.0 s, interval 1)
[hos Presentation Thread] vi: present intervals: <15.5 ms 62, 15.5-17.8 198, 17.8-22 39, 22-30 2, 30-40 0, over 40 0
[hos host] vk: replay reads of guest memory as submitted: 4260000 (2377754 pages from the captures)
[hos host] vk: a frame: 191 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 73 ms in all, 0 jobs queued
[hos host] vk: a frame's replay time (ms): submit 8.19, present 0.16, draw 4.92, descriptors 2.01, texture lookups 0.04, uploads 0.00, pipelines 0.21, programs 0.77, commands 1.58, dispatch 0.58, memory sync 0.31, queue submit 1.02, slot wait 0.06
[hos host] vk: a frame: mirrors 220 (986 MB): 5.39 MB synced in 1022.6 copies, 18.68 MB written back
[hos host] vk: 7080 presents, 60.2 fps; a frame: 2156 draws (0 skipped), 168 dispatches, 187 passes, 146 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 37379 pipelines (989 ms), 0 programs (10 failed), 4580 textures (8750 MB)
[hos host] vk: host time in GPU resource waits: 5.91 ms a frame (fences 0.00, timelines 5.91, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 666.3 full barriers recorded, 267.9 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: 42.0; 0x12: 73.5; 0x13: 2.0; 0x21: 4.0; 0x28: 14.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 11.8 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: replay reads of guest memory as submitted: 4280000 (2388118 pages from the captures)
[hos host] vk: write-backs left out (600 submits): 640, 2795.9 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2037 dirty pages from their submits' captures, 41866 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): 807 dispatches, 32862 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6399 runs, 597.0 MB compared; 64 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 14936 in captures, 1 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos ModuleSystemWorker4] nvn: CommandBufferWaitEvent #125000: event 0x100d6c8018 mode 0 value 89933 (now 89914) (lr 0x80661124)
[hos sead::MainThread] nvn: CommandBufferSignalEvent #90000: event 0x100d6c803c mode 0 value 89999 (lr 0x80661bfc)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.03
[hos host] vk: replay reads of guest memory as submitted: 4300000 (2398468 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 1866638 pages so far (0 held: may be written ahead in their submission)
[hos host] vk: replay reads of guest memory as submitted: 4320000 (2408872 pages from the captures)
[hos host] vk: shadow staging (600 takes): 0 shadow copies from staging, 19670 from the capture buffer; capture buffers: 600 host-cached, 0 not (write-combined: staged)
[hos host] vk: the recording streams: 48 of 2 MB, the most one held 347 KB; 0 made on demand (0 MB kept now); c1 stages: at most 168 a recording; 0 descriptor pools made when a recording's own ran out
[hos host] vk: sparse mirrors: 87, 986 MB with memory of 1540 MB (553 MB spared); blocks given memory later 8124, given back 14532; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 9054; 622 binds (340.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 86 taken from the spares, 134 made new; 133 kept as spares, 0 destroyed; 47 spares now, 1396 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 621, 2710.2 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2708 dirty pages from their submits' captures, 40862 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): 803 dispatches, 32291 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6272 runs, 581.5 MB compared; 57 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 15069 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 150 live (2328 MB), 0 retired still imported (0 MB), 53 retired given back (1408 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 ModuleSystemWorker5] barrier: 183353 looks: 35897 found the masks moved already, 58477 waited until they moved, 88979 for the whole 1 ms (a wait 730 us on average); 51448 wakes
[hos sead::MainThread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.04 over 667 ranges, 3853 MB), 220 pages written, 220 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.014; 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 (112 MB), 11 spare buffers, 6029 texture pages copied (23 MB)
[hos host] vk: replay reads of guest memory as submitted: 4340000 (2419209 pages from the captures)
[hos host] vk: program picks: 6827872 kept (2656472 by the last picks' textures, a uniform buffer bound since), 23316 new, 2348811 with other textures (of 9200000)
[hos host] vk: write-backs left out (600 submits): 640, 2795.9 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2391 dirty pages from their submits' captures, 41631 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): 809 dispatches, 32815 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6423 runs, 599.4 MB compared; 60 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 14538 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: replay reads of guest memory as submitted: 4360000 (2429592 pages from the captures)
[hos host] vk: replay reads of guest memory as submitted: 4380000 (2440279 pages from the captures)
[hos host] vk: texture slots registered again at another size: 400000 (of 478486 registrations; slot 31204 120x67 -> 640x360)
[hos host] vk: sparse mirrors: 87, 986 MB with memory of 1540 MB (553 MB spared); blocks given memory later 8124, given back 14532; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 9054; 622 binds (340.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 86 taken from the spares, 134 made new; 133 kept as spares, 0 destroyed; 47 spares now, 1396 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 640, 2795.9 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2765 dirty pages from their submits' captures, 40820 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): 797 dispatches, 32452 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6405 runs, 597.1 MB compared; 57 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 15030 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 150 live (2328 MB), 0 retired still imported (0 MB), 53 retired given back (1408 MB)
[hos host] vk: replay reads of guest memory as submitted: 4400000 (2450568 pages from the captures)
[hos host] vk: a submit's memory work (ms): maps 0.00, capture taken 0.02, writes collected 0.03
[hos host] vk: memory: working set 9105 MB (peak 16761), private 2534 MB, staging 256 MB in 16 arenas, 36 made so far (0 ms)
[hos host] vk: memory heaps: heap 0 (system): 15 blocks 888 MB, 39 allocations 858 MB; used 3240 of 31967 MB; heap 1 (device): 338 blocks 11282 MB, 20659 allocations 10557 MB; used 12004 of 22637 MB;
[hos host] vk: image blocks: 1 blocks 32 MB, 5 images 16 MB; 2 free ranges, the largest 15 MB
[hos host] vk: image blocks (render targets): 19 blocks 608 MB, 126 images 344 MB; 80 free ranges, the largest 31 MB
[hos host] vk: image blocks (textures, compacted): 217 blocks 6944 MB, 4438 images 6858 MB; 998 free ranges, the largest 19 MB
[hos host] vk: render target images: 573 reused from the pool, 572 made new, 440 destroyed from it; it holds 0 (0 MB)
[hos host] vk: VRAM: 12004 MB used of 22637 MB allowed (725 MB of it free in the blocks); textures 8750 MB (budget 11000 MB), 4580 textures
[hos host] vk: budget: VRAM 11282 of 22637 MB (12004 MB by the driver, 721 MB unaccounted): textures 8750, mirrors 986, blocks free 725, shadows 304, staging 256, other buffers 227, other images 31; RAM 3221 MB (the process's private memory 2534 MB): imports 2328, capture spares 732, captures 114, capture table 35, other buffers 12, others 0
[hos host] vk: textures: decoded ASTC 5782 MB, rendered 1101 MB (106 drawn in one frame, kept: 51680 KB), other 1102 MB; used in the last 8 presents 2384 MB
[hos host] vk: decoded ASTC used again: 165 textures (408 MB) copied from ones the game finalized, not decoded; 165 of those retired, 779 listed; 1192 MB of blocks hashed in 114 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 1 presents ago
[hos host] vk:   214 MB: 0x10a22aa600 1365x1365x1 x90 levels 11 NVN 0x8e ASTC, last used 1 presents ago
[hos host] vk:   64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0xa, last used 5177 presents ago
[hos host] vk:   64 MB: 0x113cba9000 512x512x64 x1 levels 1 NVN 0x3d rendered, last used 1 presents ago
[hos host] vk:   64 MB: 0x113755c000 4096x4096x1 x1 levels 1 NVN 0x34 rendered, last used 1 presents ago
[hos host] vk:   54 MB: 0x10005191c00 3072x3072x1 x3 levels 1 NVN 0x32 rendered, last used 3288 presents ago
[hos host] vk:   54 MB: 0x10004dd1a00 3072x3072x1 x3 levels 1 NVN 0x32 rendered, last used 1406 presents ago
[hos host] vk: finalized textures the GPU drew once, kept for being made again: 13 (2113 KB): 8x8x1 NVN 0x3f: 1 (2 KB); 260x260x1 NVN 0xd: 12 (2111 KB);
[hos host] vk: malloc: 15475 MB in arenas (2088 MB in use, 13386 MB free, 8 MB of it at the tops), 10 MB mapped
[hos Presentation Thread] vk: presents after the CPU's signal (600): 501 waited for it, 2.47 ms each, 0 let go after RX_VK_FRAME_GATE_MS
[hos host] vk: status: 600 frames, the slowest 30.0 ms; VRAM 11.7 of 22.1 GB
[hos host] vk: a frame: 191 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 73 ms in all, 1 jobs queued
[hos host] vk: a frame's replay time (ms): submit 8.26, present 0.16, draw 4.80, descriptors 1.93, texture lookups 0.05, uploads 0.00, pipelines 0.20, programs 0.76, commands 1.55, dispatch 0.59, memory sync 0.30, queue submit 1.06, slot wait 0.06
[hos host] vk: a frame: mirrors 220 (986 MB): 5.39 MB synced in 1023.9 copies, 18.71 MB written back
[hos host] vk: 7200 presents, 59.5 fps; a frame: 2155 draws (0 skipped), 168 dispatches, 187 passes, 146 barriers, 6 submits, 0.0 uploads (0.0 MB), 80 ub updates (56 overlays), GPU-bound wait 0.0 ms (recording slots only); 37379 pipelines (989 ms), 0 programs (10 failed), 4580 textures (8750 MB)
[hos host] vk: host time in GPU resource waits: 5.60 ms a frame (fences 0.00, timelines 5.60, idle 0.00, queries 0.00; sum across threads, not frame time; excludes pacing and driver submit/present blocking)
[hos host] vk: a frame: 666.1 full barriers recorded, 267.8 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: 42.0; 0x12: 73.5; 0x13: 2.0; 0x21: 4.0; 0x28: 14.0; 0x92: 3.0; 0x97: 1.0;
[hos host] vk: GPU time 11.7 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: queues ordered by their events: 45165 submissions stopped at a wait, 44985 went on once it was reached, 180 waits gone past; made to go by presents 0, releases 0, waits for the replay 0, 50 ms 18 (stopped 132.41 ms at most); waits met 127537, not met 0; captures at a CPU event's wait 4652, 0 of them at a wait met on arrival
[hos host] vk: the render thread (600 presents, 16.69 ms a frame): its jobs 8.77 ms a frame (submits 1.03, presents 0.16, others 7.58), idle 7.92; the game waited for room at 0 presents, 0.00 ms each (0.00 a frame; 1.77 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); 586 captures at a CPU signal; the submits' waits read in 0.053 ms a frame; readbacks: 3000 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 7.96 ms after its hand-over (20.39 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 1.79 + 0.25 (the window's copy is in the idle), after it 2.49 + 0.37; after a frame's last recording that reads or writes the game's memory come 1.01 recordings (0.02 ms of GPU time), after its last that reads it 1.01 (0.02 ms)
[hos host] vk: frame ahead (600 presents): 11.82 parts a frame recorded in the signal's job, 83.9% of them took nothing that came with the signal; of the others, 1.90 a frame took pages of the signal's capture into a mirror (86.6 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 6 of 586 frames, its first three in 6
[hos host] vk: frame ahead, what came with the signal: 586 signals with a capture, each of 112.2 pages in 9.6 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: 0.59+0.03, 0.04+0.02, 0.11+0.03, 0.19+0.11, 0.22+0.03, 0.37+0.02, 0.08+0.01, 0.28+0.03, 0.15+0.01, 0.23+0.02, 0.23+0.03, 0.00+0.01
[hos host] vk: malloc_trim: 13378 MB free in the arenas now, the process's private memory 2534 -> 2416 MB (44 ms)
[hos BaseProcExecutor2] vcores: 86084 takes, 13189 waits (mean 45 us, longest 2.4 ms, 0 over HOS_VCORE_LOG_MS), 5592 kicks, 5852 yields, 0 starved grants, 0 escapes in 5 s
[hos host] vk: replay reads of guest memory as submitted: 4420000 (2460852 pages from the captures)
[hos ModuleSystemWorker2] vk: readbacks of GPU results: 1385 of 3000 waited for their write-back, 0.43 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: 4306 kept their virtual core (nothing there to wait for), 3118 gave it up (counted by 256 a thread; RX_VK_INVALIDATE_QUICK=0: none kept)
[hos host] vk: write-backs left out (600 submits): 640, 2795.9 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2939 dirty pages from their submits' captures, 41306 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): 807 dispatches, 32830 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6433 runs, 598.9 MB compared; 110 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 14809 in captures, 0 only in the collection, 32 only by the replay itself
[hos host] vk:   read live, by pool:
[hos Presentation Thread] vk: capture: a submit 0.10 ms (lock 0.00, scan 0.04 over 669 ranges, 3853 MB), 231 pages written, 231 copied (0.9 MB)
[hos host] vk: a capture taken (ms): all 0.034; GPU wait for late pages 0.014 (1 of 600 takes), pools lock 0.000, retires and stamps 0.010, bytes copied 0.000; 24 kept (112 MB), 12 spare buffers, 6905 texture pages copied (26 MB)
[hos host] vk: replay reads of guest memory as submitted: 4440000 (2471101 pages from the captures)
[hos host] vk: recorder: 8121953 packets run in 10.0 s (busy 1.28 s, spun idle 3.71 s, slept 4865 times, 1987 of them under 200 us), 0 calls made directly (no room); the replay waited for it 30586 times, 69.20 ms in all, the longest 0.20 ms; vkEndCommandBuffer 11214/32.27 ms (max 0.20), vkResetDescriptorPool 11166/17.86 ms (max 0.01), replay_mu unlock 7738/17.75 ms (max 0.10), vmaInvalidateAllocation 319/0.92 ms (max 0.11), vkQueueSubmit2 149/0.40 ms (max 0.14)
[hos host] vk: program picks: 6975884 kept (2716232 by the last picks' textures, a uniform buffer bound since), 23318 new, 2400797 with other textures (of 9400000)
[hos ModuleSystemWorker5] nvn: DrawMeshTasksIndirect: 220000 so far (x1 0x11789a4000 x2 0x1 x3 0x11789a4000 x4 0x4040)
[hos host] vk: replay reads of guest memory as submitted: 4460000 (2481336 pages from the captures)
[hos host] vk: sparse mirrors: 87, 986 MB with memory of 1540 MB (553 MB spared); blocks given memory later 8124, given back 14532; bindings served from guest memory meanwhile 0 (0 reaching blocks with memory too), from the mirror past every game pool 9054; 622 binds (340.7 ms); 0 made whole for want of memory
[hos host] vk: sparse mirrors' buffers: 86 taken from the spares, 134 made new; 133 kept as spares, 0 destroyed; 47 spares now, 1396 MB of address space (RX_VK_SPARSE_SPARE_MB=4096)
[hos host] vk: write-backs left out (600 submits): 640, 2795.9 MB; 31 bindings followed, 47 ranges the CPU read, 0 resets; CPU reads or writes of one left out: 1686 so far
[hos host] vk: mirror syncs (600 submits): 2643 dirty pages from their submits' captures, 41462 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): 807 dispatches, 32801 chunks; 0 syncs as copies (no table room)
[hos host] vk: write-back diff (600 submits): 6391 runs, 597.8 MB compared; 58 pages read into the shadows; 304 MB of shadows; pages the shadows knew that the CPU wrote: 14999 in captures, 0 only in the collection, 31 only by the replay itself
[hos host] vk:   read live, by pool:
[hos host] vk: imports: 150 live (2328 MB), 0 retired still imported (0 MB), 53 retired given back (1408 MB)
[hos ModuleSystemWorker5] nvn: 1000000 recordings made, 57350 live copies (19.8 MB), 656 texture copies; 4 command buffers' segments (4.2 MB)
[hos host] vk: replay reads of guest memory as submitted: 4480000 (2491638 pages from the captures)
[hos host] vk: readback pages refreshed from guest memory: 1936936 pages so far (0 held: may be written ahead in their submission)
[hos sead::MainThread] nvn: GPU waits in 5 s: 252 sync waits (702 ms), 0 polls timed out (0 ms), 0 queue finishes (0 ms)
[hos host] control: quit asked by a client
[hos host] vk: replay reads of guest memory as submitted: 4500000 (2501978 pages from the captures)
[hos host] vk: indexed descriptor arrays (close): 2631 in 2631 of 55732 program stages, 0 of them near a limit
[hos host] vk: indexed descriptor arrays (close), fragment stages: 2631 texture (17-64: 2631); the longest 64
[hos host] vk: page table census (close): no pool outside its windows, 64-72 GB and 508-512 GB (4295 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)