lpc0220androot d4ed61cc7c [refactor] VSA: rename the sm100a block-sparse kernel to ptx_blackwell (#1956) ...

Co-authored-by: root <root@ptyche0124.ptyche.clusters.nvidia.com>

2026-10-17 21:10:19 -07:00

..

triton_kernels

[perf] Fuse VSA64 tile layout for SM100 inference (#1900)

2026-10-07 15:31:58 -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_bwd_ptx_blackwell.py

[refactor] VSA: rename the sm100a block-sparse kernel to ptx_blackwell (#1956)

2026-10-17 21:10:19 -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_ptx_blackwell.py

[refactor] VSA: rename the sm100a block-sparse kernel to ptx_blackwell (#1956)

2026-10-17 21:10:19 -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

[refactor] VSA: rename the sm100a block-sparse kernel to ptx_blackwell (#1956)

2026-10-17 21:10:19 -07:00

comm_ops.py

[perf] Harden Ulysses agreement and tune bounded H3 transfers (#1807)

2026-10-07 04:01:53 -07:00

ops.py

[perf] avoid full-sequence materialization in VSA coarse/sparse combine (#1813)

2026-10-07 01:54:35 -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

[refactor] VSA: rename the sm100a block-sparse kernel to ptx_blackwell (#1956)

2026-10-17 21:10:19 -07:00