Jeff Huang 908afb3a55 [FMHA] Support page_size=1 (linear layout) in batch prefill pipeline (#3545) ...

- Enable page_size=1 support in batch prefill codegen (linear layout only).
- Implement per-token page lookup in `kv_offset_array_transform` for page_size=1 to handle 3D input tensors correctly.
- Relax `kPageBlockSize` alignment assertion for the page_size=1 case.

[ROCm/composable_kernel commit: c9f112b026]

2026-01-13 12:04:43 +08:00

..

ck

moe fp8 blockscale use nt (#3524)

2026-01-12 10:48:10 +08:00

ck_tile

[FMHA] Support page_size=1 (linear layout) in batch prefill pipeline (#3545)

2026-01-13 12:04:43 +08:00

rapidjson

Update pre-commit to fixed versions, run remod for ck_tile (#2895)

2025-10-16 15:29:17 -07:00