wyj
e1ec480669
Cooperate across 32 lanes per PSF and use two completion-protected staging slots with complete batch timing. Restore coarse OpenMP event production while serializing shared GPU submissions and direct fallback boundaries. Add bounded benchmarks, streaming and renderer regressions, and preserve validation evidence and ownership documentation.
4 lines
430 B
Plaintext
4 lines
430 B
Plaintext
PSF cache ready: 64x64 phases, radius 47 px, relative tail 1e-08, tail abs 1e-06, boundary 1e-07, build 0.691 s
|
|
device=AMD Radeon RX 9070 arch=gfx1201 events=4096 spread=16 frame=3840x2160 event_bytes=56 CPU_reference=0.051886 s
|
|
HIP create=0.023212 replay_wall=0.010834 upload=0.000021 kernel=0.005702 download=0.004809 batches=1 timed=1 max_abs=7.9936057773011271e-15 max_rel=2.9244848757288279e-15 flux_rel=3.55416413758235e-17
|