Aviral Goel db79fad16f fix(grouped_gemm): pipeline selection when tail_num varies per group and leads to numerical error (#2863) ...

* fix(grouped_gemm): numerical errors on gfx950 by correctly calculating the tail num

* WIP: add temp config to stress test numerical error correction

* refactor: remove comments

2025-09-16 18:43:19 -07:00

..

block

feat(grouped_gemm): add preshuffle v2 support to grouped gemm example (#2721)

2025-09-07 14:18:35 -07:00

kernel

fix(grouped_gemm): pipeline selection when tail_num varies per group and leads to numerical error (#2863)

2025-09-16 18:43:19 -07:00

pipeline

Fix the vector load & fix the gfx950 compv4 error (#2831)

2025-09-12 11:48:45 -07:00

warp

[CK_TILE] Enable ck_tile tests on gfx11 and gfx12 (#2821)

2025-09-12 12:45:14 -07:00