Commit Graph
2 Commits
Author SHA1 Message Date
qutaoandClaude Fable 5 66c8a30c2a Pin scale layout byte-compat with CuTeDSL blocked SF layout
blockscaled_chunk_kmajor packed words are byte-identical to the
CUTLASS/CuTeDSL canonical blocked SF layout (tile_atom_to_shape_SF,
sf_vec_size=16, atom ((32,4),(16,4)):((16,4),(0,1)) tiled with order
(2,1,3)); the two views bridge zero-copy via torch view casts. Document
that on the packer and lock it with a cross-layout contract test.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-10 14:02:15 +08:00
qutao 4de007ee84 Add SM120 NVFP4 blockscaled GEMM support 2026-07-10 14:02:15 +08:00