Ding, Yi 3607588ca4 FMHA BWD workspace: 4K-align dq_acc base ...

Mark the FmhaBwdWorkspaceManager size/offset accessors as CK_TILE_HOST
(they are only invoked from host-side workspace setup), and pad
GetWorkspaceHostSize up to a 4K boundary so the GPU dq_acc buffer always
starts on a page-aligned offset.

2026-04-22 02:08:34 -05:00

..

ck

[CK][CK Tile] Conv Bwd Data flush cache and profiling improvements (#6090)

2026-04-04 00:22:22 +00:00

ck_tile

FMHA BWD workspace: 4K-align dq_acc base

2026-04-22 02:08:34 -05:00

rapidjson

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

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