add tests that are long enough to hit the vectorized path

[ghstack-poisoned]

Author: swolchok

.github/workflows/apple-perf.yml

@@ -386,8 +386,7 @@ jobs:
         echo "::endgroup::"
 
         echo "::group::Build ExecuTorch iOS frameworks"
-        PYTHON_EXECUTABLE=python ${CONDA_RUN} --no-capture-output \
-          scripts/build_apple_frameworks.sh --Release --Debug --coreml --custom --mps --optimized --portable --quantized --xnnpack
+        PYTHON_EXECUTABLE=python ${CONDA_RUN} --no-capture-output scripts/build_apple_frameworks.sh
         echo "::endgroup::"
 
         # NB: Although exported models can be copied to this directory and bundled together with the

.github/workflows/apple.yml

@@ -173,8 +173,7 @@ jobs:
         backends/apple/mps/install_requirements.sh
 
         # Build iOS Frameworks
-        PYTHON_EXECUTABLE=python ${CONDA_RUN} --no-capture-output \
-        scripts/build_apple_frameworks.sh --Release --Debug --coreml --custom --mps --optimized --portable --quantized --xnnpack
+        PYTHON_EXECUTABLE=python ${CONDA_RUN} --no-capture-output scripts/build_apple_frameworks.sh
 
         # Bundle iOS Frameworks
         for FRAMEWORK in "${FRAMEWORKS[@]}"; do (
@@ -314,8 +313,7 @@ jobs:
         echo "::endgroup::"
 
         echo "::group::Build ExecuTorch iOS frameworks"
-        PYTHON_EXECUTABLE=python ${CONDA_RUN} --no-capture-output \
-          scripts/build_apple_frameworks.sh --Release --Debug --coreml --custom --mps --optimized --portable --quantized --xnnpack
+        PYTHON_EXECUTABLE=python ${CONDA_RUN} --no-capture-output scripts/build_apple_frameworks.sh
         echo "::endgroup::"
 
         echo "::group::Build ExecuTorch benchmark app"

.github/workflows/build-presets.yml

@@ -66,3 +66,30 @@ jobs:
         ./install_requirements.sh > /dev/null
         cmake --preset ${{ matrix.preset }}
         cmake --build cmake-out -j$(( $(nproc) - 1 ))
+
+  windows:
+    uses: pytorch/test-infra/.github/workflows/windows_job.yml@main
+    strategy:
+      fail-fast: false
+      matrix:
+        preset: [pybind]
+    with:
+      job-name: build
+      ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
+      submodules: recursive
+      timeout: 90
+      script: |
+        set -eux
+        conda init powershell
+        powershell -Command "& {
+          \$ErrorActionPreference = 'Stop'
+          Set-PSDebug -Trace 1
+
+          conda create --yes --quiet -n et python=3.12
+          conda activate et
+
+          python install_requirements.py
+          cmake --preset ${{ matrix.preset }}
+          \$numCores = [System.Environment]::GetEnvironmentVariable('NUMBER_OF_PROCESSORS') - 1
+          cmake --build cmake-out -j \$numCores
+        }"

.github/workflows/trunk.yml

@@ -692,29 +692,3 @@ jobs:
       build-mode: Release
       build-tool: cmake
       docker-image: executorch-ubuntu-22.04-clang12
-
-  unittest-nxp-neutron:
-    uses: pytorch/test-infra/.github/workflows/linux_job_v2.yml@main
-    permissions:
-      id-token: write
-      contents: read
-    with:
-      runner: linux.2xlarge
-      docker-image: executorch-ubuntu-22.04-clang12
-      submodules: 'true'
-      ref: ${{ github.event_name == 'pull_request' && github.event.pull_request.head.sha || github.sha }}
-      timeout: 90
-      script: |
-        set -eux
-        
-        # The generic Linux job chooses to use base env, not the one setup by the image
-        CONDA_ENV=$(conda env list --json | jq -r ".envs | .[-1]")
-        conda activate "${CONDA_ENV}"
-        
-        # Build and install Executorch
-        PYTHON_EXECUTABLE=python \
-        CMAKE_ARGS="-DEXECUTORCH_BUILD_NXP_NEUTRON=ON" \
-        .ci/scripts/setup-linux.sh --build-tool "cmake"
-        
-        # Run pytest
-        PYTHON_EXECUTABLE=python bash backends/nxp/run_unittests.sh

.gitmodules

@@ -67,3 +67,6 @@
 [submodule "shim"]
 	path = shim
 	url = https://github.com/facebook/buck2-shims-meta
+[submodule "third-party/json"]
+	path = third-party/json
+	url = https://github.com/nlohmann/json.git

backends/apple/coreml/runtime/workspace/executorchcoreml.xcodeproj/project.pbxproj

@@ -925,7 +925,7 @@
 					"$(SRCROOT)/../include/executorch/runtime/core/portable_type/c10",
 					"$(SRCROOT)/../sdk",
 					"$(SRCROOT)/../util",
-					"$(SRCROOT)/../../third-party/nlohmann_json/single_include",
+					"$(SRCROOT)/../../../../../third-party/json/single_include",
 					"$(SRCROOT)/../../third-party/coremltools/deps/protobuf/src",
 				);
 				IPHONEOS_DEPLOYMENT_TARGET = 16.0;
@@ -957,7 +957,7 @@
 					"$(SRCROOT)/../include/executorch/runtime/core/portable_type/c10",
 					"$(SRCROOT)/../sdk",
 					"$(SRCROOT)/../util",
-					"$(SRCROOT)/../../third-party/nlohmann_json/single_include",
+					"$(SRCROOT)/../../../../../third-party/json/single_include",
 					"$(SRCROOT)/../../third-party/coremltools/deps/protobuf/src",
 				);
 				IPHONEOS_DEPLOYMENT_TARGET = 16.0;

backends/apple/coreml/scripts/install_requirements.sh

@@ -49,14 +49,6 @@ mkdir "$COREMLTOOLS_DIR_PATH/build"
 cmake -S "$COREMLTOOLS_DIR_PATH" -B "$COREMLTOOLS_DIR_PATH/build"
 cmake --build "$COREMLTOOLS_DIR_PATH/build" --parallel --target mlmodel
 
-echo "${green}ExecuTorch: Cloning nlohmann."
-git clone https://github.com/nlohmann/json.git "$COREML_DIR_PATH/third-party/nlohmann_json"
-STATUS=$?
-if [ $STATUS -ne 0 ]; then
-    echo "${red}ExecuTorch: Failed to clone nlohmann."
-    exit 1
-fi
-
 echo "${green}ExecuTorch: Copying protobuf files."
 mkdir -p "$COREML_DIR_PATH/runtime/sdk/format/"
 cp -rf "$PROTOBUF_FILES_DIR_PATH" "$COREML_DIR_PATH/runtime/sdk/format/"

backends/cadence/aot/TARGETS

@@ -388,6 +388,7 @@ python_unittest(
         "//caffe2:torch",
         "//executorch/backends/cadence/aot:compiler",
         "//executorch/backends/cadence/aot:fuse_ops",
+        "//executorch/backends/cadence/aot:graph_builder",
         "//executorch/backends/cadence/aot:ops_registrations",
         "//executorch/backends/cadence/aot:pass_utils",
         "//executorch/backends/cadence/aot:reorder_ops",

backends/cadence/aot/pass_utils.py

@@ -144,6 +144,19 @@ def nodes_not_connected_in_gm(
     return True
 
 
+# Returns the position of the first entry of a node of a given kind in the graph.
+def get_node_pos(
+    graph_module: torch.fx.GraphModule,
+    target: torch.fx.Node,
+) -> int:
+    pos = 0
+    for node in graph_module.graph.nodes:
+        if node.target == target:
+            return pos
+        pos += 1
+    return -1
+
+
 # Returns true if there is no instance of a node with target succ_target
 # positioned immediately after a node with target pred_target in the graph
 def nodes_not_adjacent_in_gm(