llama.cpp
HIP: enable WMMA-MMQ INT kernels for RDNA 3
#17576
Merged

HIP: enable WMMA-MMQ INT kernels for RDNA 3 #17576

jiachengjason
github-actions github-actions added Nvidia GPU
github-actions github-actions added ggml
jiachengjason jiachengjason marked this pull request as ready for review 15 days ago
jiachengjason jiachengjason requested a review from JohannesGaessler JohannesGaessler 15 days ago
JohannesGaessler
JohannesGaessler commented on 2025-12-01
JohannesGaessler
JohannesGaessler
JohannesGaessler approved these changes on 2025-12-03
JohannesGaessler
jiachengjason jiachengjason force pushed from d4b31101 to a34b76f4 13 days ago
jiachengjason jiachengjason force pushed from a34b76f4 to c9ec96cb 13 days ago
jiachengjason
JohannesGaessler
JohannesGaessler commented on 2025-12-04
jiachengjason enabled wmma instructions for most quantizations other than q2k
032f69d4
jiachengjason fixed the last q2_k test case failure
888b788e
jiachengjason address comments: fix out of bound write for RDNA4, add comments afte…
40e435c7
jiachengjason clean up rebase: fix ne error in half2
e4fecbca
jiachengjason fix the EditorConfig CI
685be0e1
jiachengjason jiachengjason force pushed from 59412260 to 685be0e1 12 days ago
JohannesGaessler JohannesGaessler merged 668ed765 into master 12 days ago
CISC
hjc4869
arch-btw
Beinsezii
jiachengjason
Beinsezii

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone