SolitaryThinker 76129f5a89 [perf]: add bounded H3 windows and agreed GB200 launch plans

2026-09-06 20:41:05 +00:00

..

triton_kernels

[perf] VSA Triton: widen the autotune num_stages range (the optimum was outside it) (#1706)

2026-08-26 12:30:56 -07:00

__init__.py

[misc] yapf pass under CI's interpreter (3.12) + pin hook language_version (#1702)

2026-08-08 22:24:17 -07:00

block_sparse_attn_256.py

[perf] Add FA4 CuTe backward support for VSA-256 (#1639)

2026-08-19 11:46:49 -07:00

block_sparse_attn_cute_fwd.py

[perf] Add FA4 CuTe backward support for VSA-256 (#1639)

2026-08-19 11:46:49 -07:00

block_sparse_attn_sm100a.py

[perf] Align FastH3 optimized inference profile (#1748)

2026-08-23 02:01:03 -07:00

block_sparse_attn_varlen.py

[misc] yapf pass under CI's interpreter (3.12) + pin hook language_version (#1702)

2026-08-08 22:24:17 -07:00

block_sparse_attn.py

[kernel] Route block-sparse VSA to the sm_100a forward behind FASTVIDEO_VSA_SM100A (opt-in) (#1754)

2026-08-24 15:26:56 -07:00

comm_ops.py

[perf]: add bounded H3 windows and agreed GB200 launch plans

2026-09-06 20:41:05 +00:00

ops.py

[perf] Add FA4 CuTe backward support for VSA-256 (#1639)

2026-08-19 11:46:49 -07:00

turbodiffusion_ops.py

[misc] yapf pass under CI's interpreter (3.12) + pin hook language_version (#1702)

2026-08-08 22:24:17 -07:00

version.py

[chore] release fastvideo-kernel 0.3.5 (#1775)

2026-08-27 23:18:59 -07:00

vmoba.py

[misc] yapf pass under CI's interpreter (3.12) + pin hook language_version (#1702)

2026-08-08 22:24:17 -07:00

vsa_utils.py

[kernel] sm_100a CUDA block-sparse VSA forward (Blackwell), 64- and 128-token blocks (#1719)

2026-08-20 23:16:25 -05:00