added sub packages to reduce how much CUDA stuff needs installing

.github/workflows/compile.yml

@@ -40,7 +40,7 @@ jobs:
           mkdir build
           cd build
           cmake .. ${{ matrix.defines }}
-          cmake --build . --config Release
+          cmake --build . --config Release -j $(nproc)
       - uses: actions/upload-artifact@v3
         with:
           path: ./build/libllama.so
@@ -72,7 +72,7 @@ jobs:
           mkdir build
           cd build
           cmake .. ${{ matrix.defines }}
-          cmake --build . --config Release
+          cmake --build . --config Release -j ${env:NUMBER_OF_PROCESSORS}
 
       - name: Upload artifacts
         uses: actions/upload-artifact@v3
@@ -96,10 +96,12 @@ jobs:
         with:
           repository: ggerganov/llama.cpp
 
-      - uses: Jimver/cuda-toolkit@v0.2.10
+      - uses: Jimver/cuda-toolkit@v0.2.11
         id: cuda-toolkit
         with:
           cuda: ${{ matrix.cuda }}
+          method: 'network'
+          sub-packages: '["nvcc", "cudart", "cublas", "cublas_dev", "thrust", "visual_studio_integration"]'
 
       - name: Build
         id: cmake_build