3 ms·
No. They want you to target HIP, which is compatible with both CUDA and ROCm. https://rocm.docs.amd.com/projects/HIP/en/latest/index.html https://rocm.docs.amd
by MayeulC 3y ago
No. They want you to target HIP, which is compatible with both CUDA and ROCm.
https://rocm.docs.amd.com/projects/HIP/en/latest/index.html https://rocm.docs.amd.com/projects/HIP/en/latest/index.html
- Kelteseth 3y agoAnd with HIP now supporting Windows[0] it could be a game changer. I would love to see it working with llama.cpp[1]. I tried HIP via Blender, but it did not work for my 6600XT, it still used my CPU... [0] https://github.com/RadeonOpenCompute/ROCm/issues/666#issuecomment-1655470437 https://github.com/RadeonOpenCompute/ROCm/issues/666#issueco... [1] https://github.com/ggerganov/llama.cpp/discussions/915#discussioncomment-6626936 https://github.com/ggerganov/llama.cpp/discussions/915#discu...