llama.cpp
ggml-cuda: hip: Tune stream_k algo and mmq configs to use stream_k on AMD GCN
#30022
Merged

ggml-cuda: hip: Tune stream_k algo and mmq configs to use stream_k on AMD GCN #30022

thelittlefireman
github-actions github-actions added ggml
github-actions github-actions added CUDA
thelittlefireman thelittlefireman force pushed from b8033683 to f346879f 5 days ago
thelittlefireman thelittlefireman force pushed from f346879f to 4ccfba28 5 days ago
thelittlefireman thelittlefireman marked this pull request as ready for review 4 days ago
thelittlefireman thelittlefireman requested a review 4 days ago
thelittlefireman thelittlefireman force pushed from 4ccfba28 to 96d3a1c9 4 days ago
thelittlefireman thelittlefireman force pushed from 96d3a1c9 to 64c3c6cd 3 days ago
IMbackK IMbackK assigned IMbackK IMbackK 3 days ago
thelittlefireman
ggml-gh-bot
thelittlefireman
JohannesGaessler
JohannesGaessler commented on 2026-11-17
JohannesGaessler
JohannesGaessler approved these changes on 2026-11-17
thelittlefireman ggml-cuda: hip: enable stream_k on GCN
8868d061
thelittlefireman ggml-cuda: Test on CDNA done no improvement, remove TODO
b637eee4
thelittlefireman ggml-cuda: HIP: GCN cleanup
364578cc
thelittlefireman thelittlefireman force pushed from a3bcb625 to 364578cc 1 day ago
thelittlefireman
JohannesGaessler
JohannesGaessler approved these changes on 2026-11-17
am17an
am17an approved these changes on 2026-11-17
thelittlefireman thelittlefireman merged 2e53a0ba into master 1 day ago

Login to write a write a comment.

Login via GitHub

Assignees
Labels
Milestone