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

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

JohannesGaessler merged 5 commits into ggml-org:master from jiachengjason:feat/jiachengjason/enable_mmq_kernels_for_RDNA3

github-actions added Nvidia GPU

github-actions added ggml

jiachengjason marked this pull request as ready for review 15 days ago

jiachengjason requested a review from

JohannesGaessler 15 days ago

JohannesGaessler commented on 2025-12-01

JohannesGaessler approved these changes on 2025-12-03

jiachengjason force pushed from d4b31101 to a34b76f4 13 days ago

jiachengjason force pushed from a34b76f4 to c9ec96cb 13 days ago

JohannesGaessler commented on 2025-12-04

enabled wmma instructions for most quantizations other than q2k

032f69d4

fixed the last q2_k test case failure

888b788e

address comments: fix out of bound write for RDNA4, add comments afte…

40e435c7

clean up rebase: fix ne error in half2

e4fecbca

fix the EditorConfig CI

685be0e1

jiachengjason force pushed from 59412260 to 685be0e1 12 days ago

JohannesGaessler merged 668ed765 into master 12 days ago

Reviewers

JohannesGaessler

Assignees

No one assigned

Labels

Nvidia GPU ggml

Milestone

No milestone