Document the producer-private adaptive path, the dummy support workload diagnostics, the negative result of the old serial binning and the 2026-09-13 full-day comparison (splat 282.944 s vs atomic 481.417 s). Add the corresponding raw logs, hotspot fixtures and runner evidence, and lay out the staged plan toward a sub-240 s full-day render without changing the default atomic path.
10 lines
1.3 KiB
Plaintext
10 lines
1.3 KiB
Plaintext
COMMAND OMP_NUM_THREADS=16 build/Release/replay_psf benchmarks/hip_psf_replay_2026-09-07/center_65536.events 65536 adaptive disperse 128
|
|
TIMEOUT 45s
|
|
GIT 2d01c2c101ccea8193a64bb46ceafca3009f0ef5
|
|
BINARY_SHA256 4b2e99b04076bce7856d0ca2437ec69809dc9846cc1f28eb31e9441da7fb5554
|
|
PSF cache ready: 64x64 phases, radius 47 px, relative tail 1e-08, tail abs 1e-06, boundary 1e-07, build 0.671 s
|
|
device=AMD Radeon RX 9070 arch=gfx1201 events=65536 rounds=128 frame=3840x2160 mode=adaptive distribution=synthetic-phase-preserving-dispersal contributions_per_round=432818979 CPU_reference=1.486384153 create=0.038954020
|
|
HIP PSF progress: 6651904 submitted / 6619136 completed events; 404 / 406 batches timed; kernel 4.940 s
|
|
complete_replay=6.322065115 replay_wall=6.283111095 select=0.056676626 bin=0.000000000 upload=0.021694241 prepare=0.000000000 accumulation=6.251395628 merge=0.000000000 download=0.005425636 cleanup=0.000164032 events_s=1335104.198 contributions_s=8817419980.415 center_tiles=3454208 tile_chunks=0 atomic_chunks=512 refs=0 tasks=0 scratch_used_peak=0 scratch_device_reserved=184549376 RSS_KiB=706052 max_abs=0.0028334397981558839 max_rel=127.00000000001013 flux_R=127.00000000000001 flux_G=127.00000000000001 flux_B=126.99999999999989 flux_Y=127.00000000000001 validation=skipped-for-sustained-timing
|
|
PROCESS_WALL 8.836966998 EXIT 0
|