feat(build): adjust number of parallel make jobs (#1915)

* feat(build): adjust number of parallel make jobs

* fix: update make on MacOS from brew to support --output-sync argument

* fix: cache grpc with version as part of key to improve validity of cache hits

* fix: use gmake for tests-apple to use the updated GNU make version

* fix: actually use the new make version for tests-apple

* feat: parallelize tests-extra

* feat: attempt to cache grpc build for docker images

* fix: don't quote GRPC version

* fix: don't cache go modules, we have limited cache space, better used elsewhere

* fix: release with the same version of go that we test with

* fix: don't fail on exporting cache layers

* fix: remove deprecated BUILD_GRPC docker arg from Makefile

.github/workflows/image-pr.yml

@@ -22,7 +22,7 @@ jobs:
       platforms: ${{ matrix.platforms }}
       runs-on: ${{ matrix.runs-on }}
       base-image: ${{ matrix.base-image }}
-      makeflags: "--jobs=3 --output-sync=target"
+      makeflags: ${{ matrix.makeflags }}
     secrets:
       dockerUsername: ${{ secrets.DOCKERHUB_USERNAME }}
       dockerPassword: ${{ secrets.DOCKERHUB_PASSWORD }}
@@ -42,6 +42,7 @@ jobs:
             image-type: 'extras'
             runs-on: 'arc-runner-set'
             base-image: "ubuntu:22.04"
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
             cuda-minor-version: "1"
@@ -52,6 +53,7 @@ jobs:
             image-type: 'extras'
             runs-on: 'arc-runner-set'
             base-image: "ubuntu:22.04"
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'hipblas'
             platforms: 'linux/amd64'
             tag-latest: 'false'
@@ -60,6 +62,7 @@ jobs:
             image-type: 'extras'
             base-image: "rocm/dev-ubuntu-22.04:6.0-complete"
             runs-on: 'arc-runner-set'
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'sycl_f16'
             platforms: 'linux/amd64'
             tag-latest: 'false'
@@ -68,6 +71,7 @@ jobs:
             ffmpeg: 'true'
             image-type: 'extras'
             runs-on: 'arc-runner-set'
+            makeflags: "--jobs=3 --output-sync=target"
   core-image-build:
     uses: ./.github/workflows/image_build.yml
     with:
@@ -81,7 +85,7 @@ jobs:
       platforms: ${{ matrix.platforms }}
       runs-on: ${{ matrix.runs-on }}
       base-image: ${{ matrix.base-image }}
-      makeflags: "--jobs=3 --output-sync=target"
+      makeflags: ${{ matrix.makeflags }}
     secrets:
       dockerUsername: ${{ secrets.DOCKERHUB_USERNAME }}
       dockerPassword: ${{ secrets.DOCKERHUB_PASSWORD }}
@@ -98,6 +102,7 @@ jobs:
             image-type: 'core'
             runs-on: 'ubuntu-latest'
             base-image: "ubuntu:22.04"
+            makeflags: "--jobs=5 --output-sync=target"
           - build-type: 'sycl_f16'
             platforms: 'linux/amd64'
             tag-latest: 'false'
@@ -106,6 +111,7 @@ jobs:
             ffmpeg: 'true'
             image-type: 'core'
             runs-on: 'arc-runner-set'
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
             cuda-minor-version: "1"
@@ -116,3 +122,4 @@ jobs:
             image-type: 'core'
             runs-on: 'ubuntu-latest'
             base-image: "ubuntu:22.04"
+            makeflags: "--jobs=5 --output-sync=target"

.github/workflows/image.yml

@@ -27,7 +27,7 @@ jobs:
       runs-on: ${{ matrix.runs-on }}
       base-image: ${{ matrix.base-image }}
       aio: ${{ matrix.aio }}
-      makeflags: "--jobs=3 --output-sync=target"
+      makeflags: ${{ matrix.makeflags }}
     secrets:
       dockerUsername: ${{ secrets.DOCKERHUB_USERNAME }}
       dockerPassword: ${{ secrets.DOCKERHUB_PASSWORD }}
@@ -49,6 +49,7 @@ jobs:
             image-type: 'extras'
             runs-on: 'arc-runner-set'
             base-image: "ubuntu:22.04"
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: ''
             platforms: 'linux/amd64'
             tag-latest: 'auto'
@@ -57,6 +58,7 @@ jobs:
             image-type: 'extras'
             runs-on: 'arc-runner-set'
             base-image: "ubuntu:22.04"
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "11"
             cuda-minor-version: "7"
@@ -67,6 +69,7 @@ jobs:
             image-type: 'extras'
             runs-on: 'arc-runner-set'
             base-image: "ubuntu:22.04"
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
             cuda-minor-version: "1"
@@ -77,6 +80,7 @@ jobs:
             image-type: 'extras'
             runs-on: 'arc-runner-set'
             base-image: "ubuntu:22.04"
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "11"
             cuda-minor-version: "7"
@@ -88,6 +92,7 @@ jobs:
             runs-on: 'arc-runner-set'
             base-image: "ubuntu:22.04"
             aio: "-aio-gpu-nvidia-cuda-11"
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
             cuda-minor-version: "1"
@@ -99,6 +104,7 @@ jobs:
             runs-on: 'arc-runner-set'
             base-image: "ubuntu:22.04"
             aio: "-aio-gpu-nvidia-cuda-12"
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: ''
             #platforms: 'linux/amd64,linux/arm64'
             platforms: 'linux/amd64'
@@ -108,6 +114,7 @@ jobs:
             image-type: 'extras'
             base-image: "ubuntu:22.04"
             runs-on: 'arc-runner-set'
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'hipblas'
             platforms: 'linux/amd64'
             tag-latest: 'auto'
@@ -117,6 +124,7 @@ jobs:
             aio: "-aio-gpu-hipblas"
             base-image: "rocm/dev-ubuntu-22.04:6.0-complete"
             runs-on: 'arc-runner-set'
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'hipblas'
             platforms: 'linux/amd64'
             tag-latest: 'false'
@@ -125,6 +133,7 @@ jobs:
             image-type: 'extras'
             base-image: "rocm/dev-ubuntu-22.04:6.0-complete"
             runs-on: 'arc-runner-set'
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'sycl_f16'
             platforms: 'linux/amd64'
             tag-latest: 'auto'
@@ -134,6 +143,7 @@ jobs:
             image-type: 'extras'
             runs-on: 'arc-runner-set'
             aio: "-aio-gpu-intel-f16"
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'sycl_f32'
             platforms: 'linux/amd64'
             tag-latest: 'auto'
@@ -143,6 +153,7 @@ jobs:
             image-type: 'extras'
             runs-on: 'arc-runner-set'
             aio: "-aio-gpu-intel-f32"
+            makeflags: "--jobs=3 --output-sync=target"
           # Core images
           - build-type: 'sycl_f16'
             platforms: 'linux/amd64'
@@ -152,6 +163,7 @@ jobs:
             ffmpeg: 'false'
             image-type: 'core'
             runs-on: 'arc-runner-set'
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'sycl_f32'
             platforms: 'linux/amd64'
             tag-latest: 'false'
@@ -160,6 +172,7 @@ jobs:
             ffmpeg: 'false'
             image-type: 'core'
             runs-on: 'arc-runner-set'
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'sycl_f16'
             platforms: 'linux/amd64'
             tag-latest: 'false'
@@ -168,6 +181,7 @@ jobs:
             ffmpeg: 'true'
             image-type: 'core'
             runs-on: 'arc-runner-set'
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'sycl_f32'
             platforms: 'linux/amd64'
             tag-latest: 'false'
@@ -176,6 +190,7 @@ jobs:
             ffmpeg: 'true'
             image-type: 'core'
             runs-on: 'arc-runner-set'
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'hipblas'
             platforms: 'linux/amd64'
             tag-latest: 'false'
@@ -184,6 +199,7 @@ jobs:
             image-type: 'core'
             base-image: "rocm/dev-ubuntu-22.04:6.0-complete"
             runs-on: 'arc-runner-set'
+            makeflags: "--jobs=3 --output-sync=target"
           - build-type: 'hipblas'
             platforms: 'linux/amd64'
             tag-latest: 'false'
@@ -192,6 +208,7 @@ jobs:
             image-type: 'core'
             base-image: "rocm/dev-ubuntu-22.04:6.0-complete"
             runs-on: 'arc-runner-set'
+            makeflags: "--jobs=3 --output-sync=target"
   
   core-image-build:
     uses: ./.github/workflows/image_build.yml
@@ -207,7 +224,7 @@ jobs:
       runs-on: ${{ matrix.runs-on }}
       aio: ${{ matrix.aio }}
       base-image: ${{ matrix.base-image }}
-      makeflags: "--jobs=3 --output-sync=target"
+      makeflags: ${{ matrix.makeflags }}
     secrets:
       dockerUsername: ${{ secrets.DOCKERHUB_USERNAME }}
       dockerPassword: ${{ secrets.DOCKERHUB_PASSWORD }}
@@ -225,6 +242,7 @@ jobs:
             base-image: "ubuntu:22.04"
             runs-on: 'ubuntu-latest'
             aio: "-aio-cpu"
+            makeflags: "--jobs=5 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "11"
             cuda-minor-version: "7"
@@ -235,6 +253,7 @@ jobs:
             image-type: 'core'
             base-image: "ubuntu:22.04"
             runs-on: 'ubuntu-latest'
+            makeflags: "--jobs=5 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
             cuda-minor-version: "1"
@@ -245,6 +264,7 @@ jobs:
             image-type: 'core'
             base-image: "ubuntu:22.04"
             runs-on: 'ubuntu-latest'
+            makeflags: "--jobs=5 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "11"
             cuda-minor-version: "7"
@@ -255,6 +275,7 @@ jobs:
             image-type: 'core'
             runs-on: 'ubuntu-latest'
             base-image: "ubuntu:22.04"
+            makeflags: "--jobs=5 --output-sync=target"
           - build-type: 'cublas'
             cuda-major-version: "12"
             cuda-minor-version: "1"
@@ -265,3 +286,4 @@ jobs:
             image-type: 'core'
             runs-on: 'ubuntu-latest'
             base-image: "ubuntu:22.04"
+            makeflags: "--jobs=5 --output-sync=target"

.github/workflows/image_build.yml

@@ -49,7 +49,7 @@ on:
       makeflags:
         description: 'Make Flags'
         required: false
-        default: ''
+        default: '--jobs=3 --output-sync=target'
         type: string
       aio:
         description: 'AIO Image Name'
@@ -79,6 +79,7 @@ jobs:
           && sudo apt-get install -y git
       - name: Checkout
         uses: actions/checkout@v4
+
       - name: Release space from worker
         if: inputs.runs-on == 'ubuntu-latest'
         run: |
@@ -120,6 +121,7 @@ jobs:
           sudo rm -rf "/usr/local/share/boost" || true
           sudo rm -rf "$AGENT_TOOLSDIRECTORY" || true
           df -h
+
       - name: Docker meta
         id: meta
         uses: docker/metadata-action@v5
@@ -134,6 +136,7 @@ jobs:
           flavor: |
             latest=${{ inputs.tag-latest }}
             suffix=${{ inputs.tag-suffix }}
+
       - name: Docker meta AIO (quay.io)
         if: inputs.aio != ''
         id: meta_aio
@@ -147,6 +150,7 @@ jobs:
           flavor: |
             latest=${{ inputs.tag-latest }}
             suffix=${{ inputs.aio }}
+
       - name: Docker meta AIO (dockerhub)
         if: inputs.aio != ''
         id: meta_aio_dockerhub
@@ -160,6 +164,7 @@ jobs:
           flavor: |
             latest=${{ inputs.tag-latest }}
             suffix=${{ inputs.aio }}
+
       - name: Set up QEMU
         uses: docker/setup-qemu-action@master
         with:
@@ -184,6 +189,25 @@ jobs:
           username: ${{ secrets.quayUsername }}
           password: ${{ secrets.quayPassword }}
 
+      - name: Cache GRPC
+        uses: docker/build-push-action@v5
+        with:
+          builder: ${{ steps.buildx.outputs.name }}
+          build-args: |
+            IMAGE_TYPE=${{ inputs.image-type }}
+            BASE_IMAGE=${{ inputs.base-image }}
+            MAKEFLAGS=${{ inputs.makeflags }}
+            GRPC_VERSION=v1.58.0
+          context: .
+          file: ./Dockerfile
+          cache-from: type=gha
+          cache-to: type=gha,ignore-error=true
+          target: grpc
+          platforms: ${{ inputs.platforms }}
+          push: false
+          tags: ${{ steps.meta.outputs.tags }}
+          labels: ${{ steps.meta.outputs.labels }}
+
       - name: Build and push
         uses: docker/build-push-action@v5
         with:
@@ -198,18 +222,20 @@ jobs:
             MAKEFLAGS=${{ inputs.makeflags }}
           context: .
           file: ./Dockerfile
+          cache-from: type=gha
           platforms: ${{ inputs.platforms }}
           push: ${{ github.event_name != 'pull_request' }}
           tags: ${{ steps.meta.outputs.tags }}
           labels: ${{ steps.meta.outputs.labels }}
-      -
-        name: Inspect image
+
+      - name: Inspect image
         if: github.event_name != 'pull_request'
         run: |
           docker pull localai/localai:${{ steps.meta.outputs.version }}
           docker image inspect localai/localai:${{ steps.meta.outputs.version }}
           docker pull quay.io/go-skynet/local-ai:${{ steps.meta.outputs.version }}
           docker image inspect quay.io/go-skynet/local-ai:${{ steps.meta.outputs.version }}
+
       - name: Build and push AIO image
         if: inputs.aio != ''
         uses: docker/build-push-action@v5
@@ -217,12 +243,14 @@ jobs:
           builder: ${{ steps.buildx.outputs.name }}
           build-args: |
             BASE_IMAGE=quay.io/go-skynet/local-ai:${{ steps.meta.outputs.version }}
+            MAKEFLAGS=${{ inputs.makeflags }}
           context: .
           file: ./Dockerfile.aio
           platforms: ${{ inputs.platforms }}
           push: ${{ github.event_name != 'pull_request' }}
           tags: ${{ steps.meta_aio.outputs.tags }}
           labels: ${{ steps.meta_aio.outputs.labels }}
+
       - name: Build and push AIO image (dockerhub)
         if: inputs.aio != ''
         uses: docker/build-push-action@v5
@@ -230,15 +258,18 @@ jobs:
           builder: ${{ steps.buildx.outputs.name }}
           build-args: |
             BASE_IMAGE=localai/localai:${{ steps.meta.outputs.version }}
+            MAKEFLAGS=${{ inputs.makeflags }}
           context: .
           file: ./Dockerfile.aio
           platforms: ${{ inputs.platforms }}
           push: ${{ github.event_name != 'pull_request' }}
           tags: ${{ steps.meta_aio_dockerhub.outputs.tags }}
           labels: ${{ steps.meta_aio_dockerhub.outputs.labels }}
+
       - name: job summary
         run: |
           echo "Built image: ${{ steps.meta.outputs.labels }}" >> $GITHUB_STEP_SUMMARY
+
       - name: job summary(AIO)
         if: inputs.aio != ''
         run: |

.github/workflows/release.yaml

@@ -2,6 +2,9 @@ name: Build and Release
 
 on: push
 
+env:
+  GRPC_VERSION: v1.58.0
+
 permissions:
   contents: write
 
@@ -32,7 +35,8 @@ jobs:
           submodules: true
       - uses: actions/setup-go@v4
         with:
-          go-version: '>=1.21.0'
+          go-version: '1.21.x'
+          cache: false
       - name: Dependencies
         run: |
           sudo apt-get update
@@ -54,17 +58,17 @@ jobs:
         uses: actions/cache@v3
         with:
           path: grpc
-          key: ${{ runner.os }}-grpc
+          key: ${{ runner.os }}-grpc-${{ env.GRPC_VERSION }}
       - name: Build grpc
         if: steps.cache-grpc.outputs.cache-hit != 'true'
         run: |
-          git clone --recurse-submodules -b v1.58.0 --depth 1 --shallow-submodules https://github.com/grpc/grpc && \
+          git clone --recurse-submodules -b ${{ env.GRPC_VERSION }} --depth 1 --shallow-submodules https://github.com/grpc/grpc && \
           cd grpc && mkdir -p cmake/build && cd cmake/build && cmake -DgRPC_INSTALL=ON \
             -DgRPC_BUILD_TESTS=OFF \
-            ../.. && sudo make -j12
+            ../.. && sudo make --jobs 5 --output-sync=target
       - name: Install gRPC
         run: |
-          cd grpc && cd cmake/build && sudo make -j12 install
+          cd grpc && cd cmake/build && sudo make --jobs 5 --output-sync=target install
       - name: Build
         id: build
         env:
@@ -98,7 +102,8 @@ jobs:
           submodules: true
       - uses: actions/setup-go@v4
         with:
-          go-version: '>=1.21.0'
+          go-version: '1.21.x'
+          cache: false
       - name: Dependencies
         run: |
           sudo apt-get install -y --no-install-recommends libopencv-dev
@@ -135,7 +140,8 @@ jobs:
           submodules: true
       - uses: actions/setup-go@v4
         with:
-          go-version: '>=1.21.0'
+          go-version: '1.21.x'
+          cache: false
       - name: Dependencies
         run: |
           brew install protobuf grpc

.github/workflows/test-extra.yml

@@ -40,8 +40,8 @@ jobs:
       - name: Test transformers
         run: |
            export PATH=$PATH:/opt/conda/bin
-           make -C backend/python/transformers
-           make -C backend/python/transformers test
+           make --jobs=5 --output-sync=target -C backend/python/transformers
+           make --jobs=5 --output-sync=target -C backend/python/transformers test
 
   tests-sentencetransformers:
     runs-on: ubuntu-latest
@@ -69,8 +69,8 @@ jobs:
       - name: Test sentencetransformers
         run: |
            export PATH=$PATH:/opt/conda/bin
-           make -C backend/python/sentencetransformers
-           make -C backend/python/sentencetransformers test
+           make --jobs=5 --output-sync=target -C backend/python/sentencetransformers
+           make --jobs=5 --output-sync=target -C backend/python/sentencetransformers test
 
   tests-diffusers:
     runs-on: ubuntu-latest
@@ -98,8 +98,8 @@ jobs:
       - name: Test diffusers
         run: |
            export PATH=$PATH:/opt/conda/bin
-           make -C backend/python/diffusers
-           make -C backend/python/diffusers test
+           make --jobs=5 --output-sync=target -C backend/python/diffusers
+           make --jobs=5 --output-sync=target -C backend/python/diffusers test
 
 
   tests-transformers-musicgen:
@@ -128,8 +128,8 @@ jobs:
       - name: Test transformers-musicgen
         run: |
            export PATH=$PATH:/opt/conda/bin
-           make -C backend/python/transformers-musicgen
-           make -C backend/python/transformers-musicgen test
+           make --jobs=5 --output-sync=target -C backend/python/transformers-musicgen
+           make --jobs=5 --output-sync=target -C backend/python/transformers-musicgen test
 
 
 
@@ -159,8 +159,8 @@ jobs:
       - name: Test petals
         run: |
            export PATH=$PATH:/opt/conda/bin
-           make -C backend/python/petals
-           make -C backend/python/petals test
+           make --jobs=5 --output-sync=target -C backend/python/petals
+           make --jobs=5 --output-sync=target -C backend/python/petals test
 
            
 
@@ -230,8 +230,8 @@ jobs:
   #     - name: Test bark
   #       run: |
   #          export PATH=$PATH:/opt/conda/bin
-  #          make -C backend/python/bark
-  #          make -C backend/python/bark test
+  #          make --jobs=5 --output-sync=target -C backend/python/bark
+  #          make --jobs=5 --output-sync=target -C backend/python/bark test
 
            
   # Below tests needs GPU. Commented out for now
@@ -260,8 +260,8 @@ jobs:
   #     - name: Test vllm
   #       run: |
   #          export PATH=$PATH:/opt/conda/bin
-  #          make -C backend/python/vllm
-  #          make -C backend/python/vllm test
+  #          make --jobs=5 --output-sync=target -C backend/python/vllm
+  #          make --jobs=5 --output-sync=target -C backend/python/vllm test
   tests-vallex:
     runs-on: ubuntu-latest
     steps:
@@ -286,8 +286,8 @@ jobs:
       - name: Test vall-e-x
         run: |
            export PATH=$PATH:/opt/conda/bin
-           make -C backend/python/vall-e-x
-           make -C backend/python/vall-e-x test
+           make --jobs=5 --output-sync=target -C backend/python/vall-e-x
+           make --jobs=5 --output-sync=target -C backend/python/vall-e-x test
 
   tests-coqui:
     runs-on: ubuntu-latest
@@ -313,5 +313,5 @@ jobs:
       - name: Test coqui
         run: |
            export PATH=$PATH:/opt/conda/bin
-           make -C backend/python/coqui
-           make -C backend/python/coqui test
+           make --jobs=5 --output-sync=target -C backend/python/coqui
+           make --jobs=5 --output-sync=target -C backend/python/coqui test

.github/workflows/test.yml

@@ -9,6 +9,9 @@ on:
     tags:
       - '*'
 
+env:
+  GRPC_VERSION: v1.58.0
+
 concurrency:
   group: ci-tests-${{ github.head_ref || github.ref }}-${{ github.repository }}
   cancel-in-progress: true
@@ -60,6 +63,7 @@ jobs:
         uses: actions/setup-go@v4
         with:
           go-version: ${{ matrix.go-version }}
+          cache: false
       # You can test your matrix by printing the current Go version
       - name: Display Go version
         run: go version
@@ -91,20 +95,20 @@ jobs:
         uses: actions/cache@v3
         with:
           path: grpc
-          key: ${{ runner.os }}-grpc
+          key: ${{ runner.os }}-grpc-${{ env.GRPC_VERSION }}
       - name: Build grpc
         if: steps.cache-grpc.outputs.cache-hit != 'true'
         run: |
-          git clone --recurse-submodules -b v1.58.0 --depth 1 --shallow-submodules https://github.com/grpc/grpc && \
+          git clone --recurse-submodules -b ${{ env.GRPC_VERSION }} --depth 1 --jobs 5 --shallow-submodules https://github.com/grpc/grpc && \
           cd grpc && mkdir -p cmake/build && cd cmake/build && cmake -DgRPC_INSTALL=ON \
             -DgRPC_BUILD_TESTS=OFF \
-            ../.. && sudo make -j12
+            ../.. && sudo make --jobs 5
       - name: Install gRPC
         run: |
-          cd grpc && cd cmake/build && sudo make -j12 install
+          cd grpc && cd cmake/build && sudo make --jobs 5 install
       - name: Test
         run: |
-          GO_TAGS="stablediffusion tts" make test
+          GO_TAGS="stablediffusion tts" make --jobs 5 --output-sync=target test
       - name: Setup tmate session if tests fail
         if: ${{ failure() }}
         uses: mxschmitt/action-tmate@v3
@@ -151,7 +155,7 @@ jobs:
           submodules: true
       - name: Build images
         run: |
-          docker build --build-arg FFMPEG=true --build-arg IMAGE_TYPE=core -t local-ai:tests -f Dockerfile .
+          docker build --build-arg FFMPEG=true --build-arg IMAGE_TYPE=core --build-arg MAKEFLAGS="--jobs=5 --output-sync=target" -t local-ai:tests -f Dockerfile .
           BASE_IMAGE=local-ai:tests DOCKER_AIO_IMAGE=local-ai-aio:test make docker-aio
       - name: Test
         run: |
@@ -176,17 +180,20 @@ jobs:
         uses: actions/setup-go@v4
         with:
           go-version: ${{ matrix.go-version }}
+          cache: false
       # You can test your matrix by printing the current Go version
       - name: Display Go version
         run: go version
       - name: Dependencies
         run: |
-          brew install protobuf grpc
+          brew install protobuf grpc make
       - name: Test
         run: |
           export C_INCLUDE_PATH=/usr/local/include
           export CPLUS_INCLUDE_PATH=/usr/local/include
-          BUILD_TYPE="GITHUB_CI_HAS_BROKEN_METAL" CMAKE_ARGS="-DLLAMA_F16C=OFF -DLLAMA_AVX512=OFF -DLLAMA_AVX2=OFF -DLLAMA_FMA=OFF" make test
+          # Used to run the newer GNUMake version from brew that supports --output-sync
+          export PATH="/opt/homebrew/opt/make/libexec/gnubin:$PATH"
+          BUILD_TYPE="GITHUB_CI_HAS_BROKEN_METAL" CMAKE_ARGS="-DLLAMA_F16C=OFF -DLLAMA_AVX512=OFF -DLLAMA_AVX2=OFF -DLLAMA_FMA=OFF" make --jobs 4 --output-sync=target test
       - name: Setup tmate session if tests fail
         if: ${{ failure() }}
         uses: mxschmitt/action-tmate@v3

Dockerfile

@@ -90,11 +90,35 @@ RUN if [ ! -e /usr/bin/python ]; then \
 ###################################
 ###################################
 
+FROM ${BASE_IMAGE} as grpc
+
+ARG MAKEFLAGS
+ARG GRPC_VERSION=v1.58.0
+
+ENV MAKEFLAGS=${MAKEFLAGS}
+
+WORKDIR /build
+
+RUN apt-get update && \
+    apt-get install -y g++ cmake git && \
+    apt-get clean && \
+    rm -rf /var/lib/apt/lists/*
+
+RUN git clone --recurse-submodules --jobs 4 -b ${GRPC_VERSION} --depth 1 --shallow-submodules https://github.com/grpc/grpc
+
+RUN cd grpc && \
+    mkdir -p cmake/build && \
+    cd cmake/build && \
+    cmake -DgRPC_INSTALL=ON -DgRPC_BUILD_TESTS=OFF ../.. && \
+    make
+
+###################################
+###################################
+
 FROM requirements-${IMAGE_TYPE} as builder
 
 ARG GO_TAGS="stablediffusion tts"
 ARG GRPC_BACKENDS
-ARG BUILD_GRPC=true
 ARG MAKEFLAGS
 
 ENV GRPC_BACKENDS=${GRPC_BACKENDS}
@@ -121,12 +145,9 @@ RUN if [ "${BUILD_TYPE}" = "clblas" ]; then \
 # stablediffusion does not tolerate a newer version of abseil, build it first
 RUN GRPC_BACKENDS=backend-assets/grpc/stablediffusion make build
 
-RUN if [ "${BUILD_GRPC}" = "true" ]; then \
-    git clone --recurse-submodules --jobs 4 -b v1.58.0 --depth 1 --shallow-submodules https://github.com/grpc/grpc && \
-    cd grpc && mkdir -p cmake/build && cd cmake/build && cmake -DgRPC_INSTALL=ON \
-      -DgRPC_BUILD_TESTS=OFF \
-       ../.. && make install \
-    ; fi
+COPY --from=grpc /build/grpc ./grpc/
+
+RUN cd /build/grpc/cmake/build && make install
 
 # Rebuild with defaults backends
 RUN make build
@@ -179,7 +200,7 @@ WORKDIR /build
 COPY . .
 
 COPY --from=builder /build/sources ./sources/
-COPY --from=builder /build/grpc ./grpc/
+COPY --from=grpc /build/grpc ./grpc/
 
 RUN make prepare-sources && cd /build/grpc/cmake/build && make install && rm -rf grpc
 

Makefile

@@ -355,7 +355,7 @@ prepare-e2e:
 	mkdir -p $(TEST_DIR)
 	cp -rfv $(abspath ./tests/e2e-fixtures)/gpu.yaml $(TEST_DIR)/gpu.yaml
 	test -e $(TEST_DIR)/ggllm-test-model.bin || wget -q https://huggingface.co/TheBloke/CodeLlama-7B-Instruct-GGUF/resolve/main/codellama-7b-instruct.Q2_K.gguf -O $(TEST_DIR)/ggllm-test-model.bin
-	docker build --build-arg BUILD_GRPC=true --build-arg GRPC_BACKENDS="$(GRPC_BACKENDS)" --build-arg IMAGE_TYPE=core --build-arg BUILD_TYPE=$(BUILD_TYPE) --build-arg CUDA_MAJOR_VERSION=11 --build-arg CUDA_MINOR_VERSION=7 --build-arg FFMPEG=true -t localai-tests .
+	docker build --build-arg GRPC_BACKENDS="$(GRPC_BACKENDS)" --build-arg IMAGE_TYPE=core --build-arg BUILD_TYPE=$(BUILD_TYPE) --build-arg CUDA_MAJOR_VERSION=11 --build-arg CUDA_MINOR_VERSION=7 --build-arg FFMPEG=true -t localai-tests .
 
 run-e2e-image:
 	ls -liah $(abspath ./tests/e2e-fixtures)
@@ -564,6 +564,7 @@ docker:
 		--build-arg BASE_IMAGE=$(BASE_IMAGE) \
 		--build-arg IMAGE_TYPE=$(IMAGE_TYPE) \
 		--build-arg GO_TAGS="$(GO_TAGS)" \
+		--build-arg MAKEFLAGS="$(DOCKER_MAKEFLAGS)" \
 		--build-arg BUILD_TYPE=$(BUILD_TYPE) \
 		-t $(DOCKER_IMAGE) .
 	
@@ -571,6 +572,7 @@ docker-aio:
 	@echo "Building AIO image with base $(BASE_IMAGE) as $(DOCKER_AIO_IMAGE)"
 	docker build \
 		--build-arg BASE_IMAGE=$(BASE_IMAGE) \
+		--build-arg MAKEFLAGS="$(DOCKER_MAKEFLAGS)" \
 		-t $(DOCKER_AIO_IMAGE) -f Dockerfile.aio .
 
 docker-aio-all:
@@ -582,6 +584,7 @@ docker-image-intel:
 		--build-arg BASE_IMAGE=intel/oneapi-basekit:2024.0.1-devel-ubuntu22.04 \
 		--build-arg IMAGE_TYPE=$(IMAGE_TYPE) \
 		--build-arg GO_TAGS="none" \
+		--build-arg MAKEFLAGS="$(DOCKER_MAKEFLAGS)" \
 		--build-arg BUILD_TYPE=sycl_f32 -t $(DOCKER_IMAGE) .
 
 docker-image-intel-xpu:
@@ -589,6 +592,7 @@ docker-image-intel-xpu:
 		--build-arg BASE_IMAGE=intel/oneapi-basekit:2024.0.1-devel-ubuntu22.04 \
 		--build-arg IMAGE_TYPE=$(IMAGE_TYPE) \
 		--build-arg GO_TAGS="none" \
+		--build-arg MAKEFLAGS="$(DOCKER_MAKEFLAGS)" \
 		--build-arg BUILD_TYPE=sycl_f32 -t $(DOCKER_IMAGE) .
 
 .PHONY: swagger