llama.cpp
2e53a0ba
- ggml-cuda: hip: Tune stream_k algo and mmq configs to use stream_k on AMD GCN (#30022)
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
1 day ago
ggml-cuda: hip: Tune stream_k algo and mmq configs to use stream_k on AMD GCN (#30022) * ggml-cuda: hip: enable stream_k on GCN * ggml-cuda: Test on CDNA done no improvement, remove TODO * ggml-cuda: HIP: GCN cleanup
References
#30022 - ggml-cuda: hip: Tune stream_k algo and mmq configs to use stream_k on AMD GCN
Author
thelittlefireman
Parents
744c68cd
Loading