llama.cpp
HIP/CUDA: Fix reduction in mmf getting tile widths not devisiable by 64 in mul_mat_id on wave64 devices
#30836
Merged
Merged
HIP/CUDA: Fix reduction in mmf getting tile widths not devisiable by 64 in mul_mat_id on wave64 devices #30836
ggerganov
merged
1 commit into
ggml-org:master
from
IMbackK:mmid_w64
HIP/CUDA: Fix reduction in mmf getting tile widths not devisiable by …
c295a3d9
thelittlefireman
approved these changes
on 2026-10-19
Reviewers
Assignees
No one assigned
Labels
ggml
merge ready
CUDA
Milestone
No milestone
Login to write a write a comment.
Login via GitHub