features/bladedisc rebase 20220830

.github/actions/setup-build/action.yml

@@ -11,26 +11,29 @@ inputs:
 
 runs:
   using: "composite"
+
   steps:
   - name: Set up Python
     uses: actions/setup-python@v2
     with:
       python-version: 3.9
+
   - name: Install MLIR Python depends
     run: |
       python -m pip install -r $GITHUB_WORKSPACE/externals/llvm-project/mlir/python/requirements.txt
     shell: bash
+
   - name: Install PyTorch nightly depends
     run: |
       python -m pip install -r requirements.txt
     shell: bash
+
   - name: Install Ninja
     uses: llvm/actions/install-ninja@55d844821959226fab4911f96f37071c1d4c3268
-  - name: Get Submodule Hash
-    id: get-submodule-hash
-    run: echo "::set-output name=hash::$(md5sum $(git submodule status))"
-    shell: bash
+
   - name: Ccache for C++ compilation
-    uses: hendrikmuhs/ccache-action@4687d037e4d7cf725512d9b819137a3af34d39b3
+    uses: hendrikmuhs/ccache-action@v1.2
     with:
-      key: ${{ runner.os }}-clangreleaseasserts-${{ steps.get-submodule-hash.outputs.hash }}${{ inputs.cache-suffix }}
+      key: ${{ runner.os }}-torch_mlir_build_assets-${{ inputs.cache-suffix }}
+      max-size: 2G
+      verbose: 2

.github/workflows/bazelBuildAndTest.yml

@@ -2,26 +2,68 @@ name: Bazel Build and Test
 
 on:
   push:
-    branches:
-      - main
+    branches: [ main ]
+  workflow_dispatch:
+
+# Ensure that only a single job or workflow using the same
+# concurrency group will run at a time. This would cancel
+# any in-progress jobs in the same github workflow and github
+# ref (e.g. refs/heads/main or refs/pull/<pr_number>/merge).
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
 
 jobs:
-  build:
-    name: Build and Test (Release Asserts)
-    runs-on: ubuntu-20.04
+  ubuntu-build:
+    name: ubuntu-x86_64
+    runs-on: ubuntu-22.04
+
     steps:
-    - name: Set up Python
-      uses: actions/setup-python@v2
-      with:
-        python-version: 3.9
-    - name: Get torch-mlir
-      uses: actions/checkout@v2
+    - name: Checkout torch-mlir
+      uses: actions/checkout@v3
       with:
         submodules: 'true'
-    - name: Build with bazel
+
+    - name: Setup cache for bazel
+      uses: actions/cache@v3
+      with:
+        path: ~/.cache/bazel
+        key: ubuntu_x86_64_torch_mlir_bazel_build_cache
+
+    # Change bazel cache directory to root ownership
+    # to allow writing to it from within the docker container.
+    # If no cache hits, this directory is not present
+    # so don't run chown (will error otherwise).
+    - name: Set bazel cache permissions
       run: |
-        cd $GITHUB_WORKSPACE/utils/bazel
-        bazel build @torch-mlir//...
+        if [ -d "${HOME}/.cache/bazel" ]; then
+          sudo chown -R root:root "${HOME}/.cache/bazel"
+        fi
+
+    - name: Build docker image
+      run: |
+        docker build -f utils/bazel/docker/Dockerfile \
+                     -t torch-mlir:ci \
+                     .
+
+    - name: Bazel build torch-mlir
+      run: |
+        docker run --rm \
+                   -v "$(pwd)":"/opt/src/torch-mlir" \
+                   -v "${HOME}/.cache/bazel":"/root/.cache/bazel" \
+                   torch-mlir:ci \
+                   ./utils/bazel/docker/run_bazel_build.sh
+
+    # Switch back bazel cache directory to user ownership
+    # to allow GHA post-cache step to save cache without
+    # permissions issue.
+    - name: Switch bazel cache permissions
+      run: |
+        if [ -d "${HOME}/.cache/bazel" ]; then
+          sudo chown -R "$USER":"$USER" "${HOME}/.cache/bazel"
+        fi
+
     - name: Send mail
       if: failure()
       uses: dawidd6/action-send-mail@v3

.github/workflows/buildAndTest.yml

@@ -1,102 +1,201 @@
 name: Build and Test
 
 on:
-  push:
-    branches:
-      - main
   pull_request:
+    branches: [ main ]
+  push:
+    branches: [ main ]
   workflow_dispatch:
 
+# Ensure that only a single job or workflow using the same
+# concurrency group will run at a time. This would cancel
+# any in-progress jobs in the same github workflow and github
+# ref (e.g. refs/heads/main or refs/pull/<pr_number>/merge).
+concurrency:
+  group: ${{ github.workflow }}-${{ github.ref }}
+  cancel-in-progress: true
+
+
+# Provisioned Jobs:
+# ubuntu - x86_64 - llvm in-tree     - pytorch binary - build+test    # most used dev flow and fastest signal
+# ubuntu - x86_64 - llvm out-of-tree - pytorch source - build+test    # most elaborate build
+# macos  - arm64  - llvm in-tree     - pytorch binary - build only    # cross compile, can't test arm64
 jobs:
-  build:
-    name: Build and Test (Release Asserts)
-    # Changes to the name of this job needs to be synced with releaseSnapshotPackage.yml.
-    runs-on: ubuntu-20.04
+  build-test:
+    strategy:
+      fail-fast: true
+      matrix:
+        os-arch: [ubuntu-x86_64, macos-arm64]
+        llvm-build: [in-tree, out-of-tree]
+        torch-binary: [ON, OFF]
+        exclude:
+          # Exclude llvm in-tree and pytorch source
+          - llvm-build: in-tree
+            torch-binary: OFF
+          # Exclude llvm out-of-tree and pytorch binary
+          - llvm-build: out-of-tree
+            torch-binary: ON
+          # Exclude macos-arm64 and llvm out-of-tree altogether
+          - os-arch: macos-arm64
+            llvm-build: out-of-tree
+        include:
+          # Specify OS versions
+          - os-arch: ubuntu-x86_64
+            os: ubuntu-22.04
+          - os-arch: macos-arm64
+            os: macos-12
+    runs-on: ${{ matrix.os }}
+
     steps:
-    - name: Get torch-mlir
+    - name: Checkout torch-mlir
       uses: actions/checkout@v2
       with:
         submodules: 'true'
-    - uses: ./.github/actions/setup-build
+
+    - name: Setup ccache
+      uses: ./.github/actions/setup-build
       with:
-        cache-suffix: ''
-    - name: Build and Test torch-mlir (Assert)
+        cache-suffix: ${{ matrix.os-arch }}-${{ matrix.llvm-build }}-${{ matrix.torch-binary }}
+
+    - name: Configure os-arch='ubuntu-x86_64' llvm-build='in-tree' torch-binary='${{ matrix.torch-binary }}'
+      # Fastest build, most used dev flow
+      if: ${{ matrix.os-arch == 'ubuntu-x86_64' && matrix.llvm-build == 'in-tree' }}
       run: |
-        cd $GITHUB_WORKSPACE
-        mkdir build
-        cd build
-        cmake $GITHUB_WORKSPACE/externals/llvm-project/llvm -GNinja \
+        cmake -GNinja -Bbuild \
           -DCMAKE_BUILD_TYPE=Release \
+          -DCMAKE_C_COMPILER=clang \
+          -DCMAKE_CXX_COMPILER=clang++ \
+          -DCMAKE_C_COMPILER_LAUNCHER=ccache \
+          -DCMAKE_CXX_COMPILER_LAUNCHER=ccache \
           -DCMAKE_LINKER=lld \
-          -DCMAKE_C_COMPILER_LAUNCHER=ccache -DCMAKE_CXX_COMPILER_LAUNCHER=ccache \
-          -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++ \
-          -DPython3_EXECUTABLE=$(which python) \
           -DLLVM_ENABLE_ASSERTIONS=ON \
           -DLLVM_ENABLE_PROJECTS=mlir \
           -DLLVM_EXTERNAL_PROJECTS="torch-mlir;torch-mlir-dialects" \
           -DLLVM_EXTERNAL_TORCH_MLIR_SOURCE_DIR="$GITHUB_WORKSPACE" \
-          -DLLVM_EXTERNAL_TORCH_MLIR_DIALECTS_SOURCE_DIR="${GITHUB_WORKSPACE}/external/llvm-external-projects/torch-mlir-dialects" \
+          -DLLVM_EXTERNAL_TORCH_MLIR_DIALECTS_SOURCE_DIR="${GITHUB_WORKSPACE}/externals/llvm-external-projects/torch-mlir-dialects" \
+          -DLLVM_TARGETS_TO_BUILD=host \
           -DMLIR_ENABLE_BINDINGS_PYTHON=ON \
-          -DLLVM_TARGETS_TO_BUILD=host
-        ninja check-torch-mlir-all
-    - name: RefBackend - TorchScript end-to-end tests
-      run: |
-        cd $GITHUB_WORKSPACE
-        export PYTHONPATH="$GITHUB_WORKSPACE/build/tools/torch-mlir/python_packages/torch_mlir"
-        python -m e2e_testing.torchscript.main --config=refbackend -v
-    - name: EagerMode - TorchScript end-to-end tests
-      run: |
-        cd $GITHUB_WORKSPACE
-        export PYTHONPATH="$GITHUB_WORKSPACE/build/tools/torch-mlir/python_packages/torch_mlir"
-        python -m e2e_testing.torchscript.main --config=eager_mode -v
-    - name: TOSA backend - TorchScript end-to-end tests
-      run: |
-        cd $GITHUB_WORKSPACE
-        export PYTHONPATH="$GITHUB_WORKSPACE/build/tools/torch-mlir/python_packages/torch_mlir"
-        python -m e2e_testing.torchscript.main --config=tosa -v
+          -DTORCH_MLIR_USE_INSTALLED_PYTORCH="${{ matrix.torch-binary }}" \
+          -DPython3_EXECUTABLE="$(which python)" \
+          $GITHUB_WORKSPACE/externals/llvm-project/llvm
 
-  build-out-of-tree:
-    name: Build out-of-tree (Release Asserts)
-    runs-on: ubuntu-20.04
-    steps:
-    - name: Get torch-mlir
-      uses: actions/checkout@v2
-      with:
-        submodules: 'true'
-    - uses: ./.github/actions/setup-build
-      with:
-        cache-suffix: '-out-of-tree'
-    - name: Build LLVM (standalone)
-      # This build takes a while but is expected to almost always be cached.
-      # A cache invalidation occurs when the committed LLVM version is changed.
+    - name: Configure os-arch='ubuntu-x86_64' llvm-build='out-of-tree' torch-binary='${{ matrix.torch-binary }}'
+      # Most elaborate build, but cached
+      if: ${{ matrix.os-arch == 'ubuntu-x86_64' && matrix.llvm-build == 'out-of-tree' }}
       run: |
-        cd $GITHUB_WORKSPACE
-        cmake -Bllvm-build -GNinja \
+        cmake -GNinja -Bllvm-build \
           -DCMAKE_BUILD_TYPE=Release \
+          -DCMAKE_C_COMPILER=clang \
+          -DCMAKE_CXX_COMPILER=clang++ \
+          -DCMAKE_C_COMPILER_LAUNCHER=ccache \
+          -DCMAKE_CXX_COMPILER_LAUNCHER=ccache \
           -DCMAKE_LINKER=lld \
-          -DCMAKE_C_COMPILER_LAUNCHER=ccache -DCMAKE_CXX_COMPILER_LAUNCHER=ccache \
-          -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++ \
-          -DPython3_EXECUTABLE=$(which python) \
           -DLLVM_ENABLE_ASSERTIONS=ON \
           -DLLVM_ENABLE_PROJECTS=mlir \
-          -DMLIR_ENABLE_BINDINGS_PYTHON=ON \
           -DLLVM_TARGETS_TO_BUILD=host \
-          externals/llvm-project/llvm
-        ninja -Cllvm-build
+          -DMLIR_ENABLE_BINDINGS_PYTHON=ON \
+          -DPython3_EXECUTABLE="$(which python)" \
+          $GITHUB_WORKSPACE/externals/llvm-project/llvm
+        cmake --build llvm-build
 
-    - name: Build and test torch-mlir (out-of-tree)
-      run: |
-        cd $GITHUB_WORKSPACE
         cmake -GNinja -Bbuild \
+          -DCMAKE_C_COMPILER=clang \
+          -DCMAKE_CXX_COMPILER=clang++ \
+          -DCMAKE_C_COMPILER_LAUNCHER=ccache \
+          -DCMAKE_CXX_COMPILER_LAUNCHER=ccache \
+          -DCMAKE_LINKER=lld \
+          -DLLVM_DIR="$GITHUB_WORKSPACE/llvm-build/lib/cmake/llvm/" \
+          -DMLIR_DIR="$GITHUB_WORKSPACE/llvm-build/lib/cmake/mlir/" \
+          -DMLIR_ENABLE_BINDINGS_PYTHON=OFF \
+          -DTORCH_MLIR_USE_INSTALLED_PYTORCH="${{ matrix.torch-binary }}" \
+          -DPython3_EXECUTABLE="$(which python)" \
+          $GITHUB_WORKSPACE
+
+    - name: Configure os-arch='macos-arm64' llvm-build='in-tree' torch-binary='${{ matrix.torch-binary }}'
+      # cross compile, can't test arm64
+      if: ${{ matrix.os-arch == 'macos-arm64' && matrix.llvm-build == 'in-tree' }}
+      run: |
+        # TODO: Reenable LTC after build on macOS-arm64 is fixed (https://github.com/llvm/torch-mlir/issues/1253)
+        cmake -GNinja -Bbuild_arm64 \
+          -DCMAKE_BUILD_TYPE=Release \
+          -DCMAKE_C_COMPILER=clang \
+          -DCMAKE_CXX_COMPILER=clang++ \
+          -DCMAKE_C_COMPILER_LAUNCHER=ccache \
+          -DCMAKE_CXX_COMPILER_LAUNCHER=ccache \
           -DCMAKE_LINKER=lld \
-          -DCMAKE_C_COMPILER_LAUNCHER=ccache -DCMAKE_CXX_COMPILER_LAUNCHER=ccache \
-          -DCMAKE_C_COMPILER=clang -DCMAKE_CXX_COMPILER=clang++ \
-          -DMLIR_DIR="$(pwd)/llvm-build/lib/cmake/mlir/" \
-          -DLLVM_DIR="$(pwd)/llvm-build/lib/cmake/llvm/" \
+          -DCMAKE_OSX_ARCHITECTURES=arm64 \
+          -DLLVM_ENABLE_ASSERTIONS=ON \
+          -DLLVM_ENABLE_PROJECTS=mlir \
+          -DLLVM_EXTERNAL_PROJECTS="torch-mlir;torch-mlir-dialects" \
+          -DLLVM_EXTERNAL_TORCH_MLIR_SOURCE_DIR="$GITHUB_WORKSPACE" \
+          -DLLVM_EXTERNAL_TORCH_MLIR_DIALECTS_SOURCE_DIR="${GITHUB_WORKSPACE}/externals/llvm-external-projects/torch-mlir-dialects" \
+          -DLLVM_TARGETS_TO_BUILD=AArch64 \
+          -DLLVM_USE_HOST_TOOLS=ON \
+          -DLLVM_ENABLE_ZSTD=OFF \
           -DMLIR_ENABLE_BINDINGS_PYTHON=ON \
-          -DPython3_EXECUTABLE=$(which python) \
-          .
-        ninja -Cbuild check-torch-mlir-all
+          -DTORCH_MLIR_ENABLE_MHLO=OFF \
+          -DTORCH_MLIR_ENABLE_LTC=OFF \
+          -DTORCH_MLIR_USE_INSTALLED_PYTORCH="${{ matrix.torch-binary }}" \
+          -DMACOSX_DEPLOYMENT_TARGET=12.0 \
+          -DPython3_EXECUTABLE="$(which python)" \
+          $GITHUB_WORKSPACE/externals/llvm-project/llvm
+
+    - name: Build torch-mlir
+      if: ${{ matrix.os-arch == 'ubuntu-x86_64' }}
+      run: |
+        cmake --build build
+
+    - name: Build torch-mlir (cross-compile)
+      if: ${{ matrix.os-arch == 'macos-arm64' }}
+      run: |
+        cmake --build build_arm64
+
+    - name: Run torch-mlir unit tests
+      if: ${{ matrix.os-arch == 'ubuntu-x86_64' }}
+      run: |
+        cmake --build build --target check-torch-mlir-all
+
+    - name: Ensure generated files are up to date
+      if: ${{ matrix.os-arch == 'ubuntu-x86_64' && matrix.llvm-build == 'in-tree' }}
+      run: |
+        ./build_tools/update_torch_ods.sh
+        ./build_tools/update_shape_lib.sh
+        if ! git diff --quiet; then
+          echo "#######################################################"
+          echo "Generated files are not up to date (see diff below)"
+          echo ">>> Please run ./build_tools/update_torch_ods.sh and ./build_tools/update_shape_lib.sh <<<"
+          echo "#######################################################"
+          git diff --color=always
+          exit 1
+        fi
+
+    - name: Run refbackend e2e integration tests
+      if: ${{ matrix.os-arch == 'ubuntu-x86_64' && matrix.llvm-build == 'in-tree' }}
+      run: |
+        export PYTHONPATH="$GITHUB_WORKSPACE/build/tools/torch-mlir/python_packages/torch_mlir"
+        python -m e2e_testing.main --config=refbackend -v
 
-    # Don't run python tests, as check-torch-mlir-all already checks
-    # what we want.
+    - name: Run eager_mode e2e integration tests
+      if: ${{ matrix.os-arch == 'ubuntu-x86_64' && matrix.llvm-build == 'in-tree' }}
+      run: |
+        export PYTHONPATH="$GITHUB_WORKSPACE/build/tools/torch-mlir/python_packages/torch_mlir"
+        python -m e2e_testing.main --config=eager_mode -v
+
+    - name: Run mhlo e2e integration tests
+      if: ${{ matrix.os-arch == 'ubuntu-x86_64' && matrix.llvm-build == 'in-tree' }}
+      run: |
+        export PYTHONPATH="$GITHUB_WORKSPACE/build/tools/torch-mlir/python_packages/torch_mlir"
+        python -m e2e_testing.main --config=mhlo -v
+
+    - name: Run tosa e2e integration tests
+      if: ${{ matrix.os-arch == 'ubuntu-x86_64' && matrix.llvm-build == 'in-tree' }}
+      run: |
+        export PYTHONPATH="$GITHUB_WORKSPACE/build/tools/torch-mlir/python_packages/torch_mlir"
+        python -m e2e_testing.main --config=tosa -v
+
+    - name: Run lazy_tensor_core e2e integration tests
+      if: ${{ matrix.os-arch == 'ubuntu-x86_64' && matrix.llvm-build == 'in-tree' }}
+      run: |
+        export PYTHONPATH="$GITHUB_WORKSPACE/build/tools/torch-mlir/python_packages/torch_mlir"
+        echo "LTC tests disabled temporarily. https://github.com/llvm/torch-mlir/pull/1292"
+        # python -m e2e_testing.main --config=lazy_tensor_core -v

.github/workflows/buildRelease.yml

@@ -54,7 +54,7 @@ jobs:
 
   build_macos:
     name: MacOS Build
-    runs-on: macos-latest
+    runs-on: macos-12
     steps:
     - name: Get torch-mlir
       uses: actions/checkout@v2