highperfocused/ollama
1
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2025-08-09 06:22:41 +02:00
Code Issues Packages Projects Releases Wiki Activity

metal: add missing barriers for mul-mat (#469)

Browse Source
This commit is contained in:
Bruce MacDonald
2023-09-05 19:37:13 -04:00
committed by GitHub
parent 9ae76ba8c9
commit d18282bfda
6 changed files with 111 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
llm/llama.cpp/generate_darwin_arm64.go
View File

@@ -4,5 +4,7 @@ package llm
//go:generate git submodule update --force ggml
//go:generate git -C ggml apply ../ggml_patch/0001-add-detokenize-endpoint.patch
//go:generate git -C ggml apply ../ggml_patch/0002-34B-model-support.patch
//go:generate cmake -S ggml -B ggml/build/gpu -DLLAMA_METAL=on -DLLAMA_ACCELERATE=on -DLLAMA_K_QUANTS=on -DCMAKE_SYSTEM_PROCESSOR=arm64 -DCMAKE_OSX_ARCHITECTURES=arm64
//go:generate git -C ggml apply ../ggml_patch/0003-metal-fix-synchronization-in-new-matrix-multiplicati.patch
//go:generate git -C ggml apply ../ggml_patch/0004-metal-add-missing-barriers-for-mul-mat-2699.patch
//go:generate cmake --fresh -S ggml -B ggml/build/gpu -DLLAMA_METAL=on -DLLAMA_ACCELERATE=on -DLLAMA_K_QUANTS=on -DCMAKE_SYSTEM_PROCESSOR=arm64 -DCMAKE_OSX_ARCHITECTURES=arm64
//go:generate cmake --build ggml/build/gpu --target server --config Release