Skip to content

release: ship CPU backend variants in the CUDA and HIP bundles (#241) #271

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
khosravipasha merged 1 commit into from
Sep 25, 2026
Merged

release: ship CPU backend variants in the CUDA and HIP bundles (#241) #271

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 8 additions & 1 deletion .github/workflows/release-prism.yml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -178,6 +178,8 @@ jobs:
-DCMAKE_INSTALL_RPATH='$ORIGIN' \
-DCMAKE_BUILD_WITH_INSTALL_RPATH=ON \
-DGGML_NATIVE=OFF \
-DGGML_BACKEND_DL=ON \
-DGGML_CPU_ALL_VARIANTS=ON \
-DGGML_CUDA=ON \
${{ env.CMAKE_ARGS }}
cmake --build build --config Release -j $(nproc) 2>&1 | grep -v "^nvcc warning"
Expand Down Expand Up @@ -301,10 +303,14 @@ jobs:
call "C:\Program Files\Microsoft Visual Studio\2022\Enterprise\VC\Auxiliary\Build\vcvarsall.bat" ${{ matrix.arch == 'x64' && 'x64' || 'amd64_arm64' }}
rem arm64: ggml-cpu refuses MSVC for ARM (needs clang), so that leg builds only the
rem ggml-cuda.dll companion for the BACKEND_DL cpu-arm64 zip, like upstream does.
rem x64: BACKEND_DL + CPU_ALL_VARIANTS so the bundled CPU backend is picked at runtime.
rem GGML_NATIVE=OFF alone compiles one ggml-cpu.dll with AVX2/FMA, which dies silently
rem at startup on CPUs without AVX2 (#241).
Comment thread
khosravipasha marked this conversation as resolved.
Resolved
cmake -S . -B build -G "Ninja Multi-Config" ^
-DGGML_NATIVE=OFF ^
-DGGML_CUDA=ON ^
${{ matrix.arch == 'arm64' && '-DGGML_BACKEND_DL=ON -DGGML_CPU=OFF' || '' }} ^
-DGGML_BACKEND_DL=ON ^
${{ matrix.arch == 'arm64' && '-DGGML_CPU=OFF' || '-DGGML_CPU_ALL_VARIANTS=ON' }} ^
-DLLAMA_BUILD_BORINGSSL=ON ^
-DCMAKE_CUDA_FLAGS="-diag-suppress=221" ^
${{ matrix.defines }} ${{ env.CMAKE_ARGS }}
Expand Down Expand Up @@ -623,6 +629,7 @@ jobs:
-DCMAKE_BUILD_TYPE=Release `
-DGGML_BACKEND_DL=ON `
-DGGML_NATIVE=OFF `
-DGGML_CPU_ALL_VARIANTS=ON `
-DGGML_OPENMP=OFF `
-DGPU_TARGETS="${{ matrix.gpu_targets }}" `
-DGGML_HIP_ROCWMMA_FATTN=ON `
Expand Down
Loading