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.
9 lines
1.2 KiB
Plaintext
9 lines
1.2 KiB
Plaintext
COMMAND OMP_NUM_THREADS=16 OMP_DYNAMIC=FALSE env HIP_PSF_ACCUMULATION=adaptive build/Release/replay_psf benchmarks/hip_psf_replay_2026-09-07/center_65536.events 65536 production 128
|
|
TIMEOUT 45s
|
|
GIT 2d01c2c101ccea8193a64bb46ceafca3009f0ef5
|
|
BINARY_SHA256 61b02b48a9b16fdb2f82ab7d6ce9d2d96e8c964a30b2243761b45c21ea346808
|
|
PSF cache ready: 64x64 phases, radius 47 px, relative tail 1e-08, tail abs 1e-06, boundary 1e-07, build 0.693 s
|
|
device=AMD Radeon RX 9070 arch=gfx1201 events=65536 rounds=128 frame=3840x2160 mode=production distribution=captured contributions_per_round=432818979 CPU_reference=0.829993010 create=0.041409016
|
|
complete_replay=3.920389891 replay_wall=3.878980875 select=0.022354766 bin=0.552004015 upload=0.085706228 prepare=0.000000000 accumulation=3.654319396 merge=0.000000000 download=0.005400386 cleanup=0.000000000 events_s=2162580.397 contributions_s=14282315664.107 center_tiles=0 tile_chunks=512 atomic_chunks=0 refs=0 tasks=0 scratch_used_peak=0 scratch_device_reserved=0 RSS_KiB=715700 max_abs=0.0079141092811036706 max_rel=127.00000000000276 flux_R=126.99999999999998 flux_G=127.00000000000002 flux_B=127.00000000000001 flux_Y=127 validation=skipped-for-sustained-timing
|
|
PROCESS_WALL 5.824096583 EXIT 0
|