Sync with Microsoft ONNX Runtime - 17/09/2025

.github/workflows/android.yml

@@ -71,8 +71,8 @@ jobs:
       run: |
         set -e -x
         BINARY_SIZE_THRESHOLD_ARGS=""
-        echo "Binary size threshold in bytes: 1436672"
-        BINARY_SIZE_THRESHOLD_ARGS="--threshold_size_in_bytes 1436672"
+        echo "Binary size threshold in bytes: 1722565"
+        BINARY_SIZE_THRESHOLD_ARGS="--threshold_size_in_bytes 1722565"
 
         # Ensure ANDROID_NDK_HOME is available and get its real path
         if [ -z "$ANDROID_NDK_HOME" ]; then

.github/workflows/linux-wasm-ci-build-and-test-workflow.yml

@@ -120,7 +120,7 @@ jobs:
           path: ${{ github.workspace }}/artifacts/wasm
 
       - name: Test (Node.js) (simd + threads)
-        # onnxruntime_test_all is currently only supported in Debug build because it requires exception, which is disabled in Release build.
+        # unit tests are currently only supported in Debug builds because they require exceptions, which are disabled in Release builds.
         if: ${{ inputs.build_config == 'Debug' }}
         run: |
           python ./tools/ci_build/build.py \
@@ -130,14 +130,14 @@ jobs:
         working-directory: ${{ github.workspace }}
 
       - name: Test (browser) (simd + threads)
-        # onnxruntime_test_all is currently only supported in Debug build because it requires exception, which is disabled in Release build.
+        # unit tests are currently only supported in Debug builds because they require exceptions, which are disabled in Release builds.
         if: ${{ inputs.build_config == 'Debug' }}
         run: |
           python ./tools/ci_build/build.py \
             ${{ env.common_build_args }} \
             --build_dir ${{ github.workspace }}/build/wasm_inferencing \
             --wasm_run_tests_in_browser \
-            --target onnxruntime_test_all \
+            --targets onnxruntime_test_all onnxruntime_provider_test \
             --update --build --test
         working-directory: ${{ github.workspace }}
 

.github/workflows/windows_qnn_x64.yml

@@ -0,0 +1,82 @@
+name: Windows x64 QNN CI Pipeline
+
+on:
+  push:
+    branches:
+      - main
+      - rel-*
+  pull_request:
+    branches:
+      - main
+      - rel-*
+  workflow_dispatch:
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event_name == 'pull_request' && github.ref || github.sha }}
+  cancel-in-progress: true
+
+jobs:
+  build_test_qnn_ep:
+    name: Windows x64 QNN CI Pipeline (${{ matrix.QnnLibKind }})
+    runs-on: ["self-hosted", "1ES.Pool=onnxruntime-github-vs2022-mms"]
+    timeout-minutes: 120
+    strategy:
+      matrix: 
+        QnnLibKind: [shared_lib, static_lib]
+    env:
+      AZCOPY_AUTO_LOGIN_TYPE: MSI
+      AZCOPY_MSI_CLIENT_ID: 63b63039-6328-442f-954b-5a64d124e5b4
+      DOTNET_SKIP_FIRST_TIME_EXPERIENCE: true
+      ALLOW_RELEASED_ONNX_OPSET_ONLY: '1'
+
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v5
+
+      - name: Setup Python
+        uses: actions/setup-python@v5
+        with:
+          python-version: '3.12'
+          architecture: x64
+
+      - name: Locate vcvarsall and Setup Env
+        uses: ./.github/actions/locate-vcvarsall-and-setup-env
+        with:
+          architecture: x64
+
+      - name: Download QNN SDK
+        working-directory: ${{ runner.temp }}
+        run: |
+          azcopy.exe cp --recursive https://lotusscus.blob.core.windows.net/models/qnnsdk/qnn-v2.38.0.250901 .
+          dir
+        shell: pwsh
+
+      - name: Set QNN_SDK_ROOT environment variable
+        shell: pwsh
+        run: |
+          $qnn_sdk_path = Join-Path $env:RUNNER_TEMP "qnn-v2.38.0.250901"
+          echo "QNN_SDK_ROOT=$qnn_sdk_path" | Out-File -FilePath $env:GITHUB_ENV -Encoding utf8 -Append
+          echo "QNN SDK Root: $qnn_sdk_path"
+          dir $qnn_sdk_path
+
+      - name: Build and Test
+        shell: cmd
+        run: |
+          python ${{ github.workspace }}\tools\ci_build\build.py --config RelWithDebInfo --build_dir ${{ runner.temp }}\build --cmake_generator "Visual Studio 17 2022" --build_java --build_shared_lib --use_qnn ${{ matrix.QnnLibKind }} --qnn_home %QNN_SDK_ROOT% --use_binskim_compliant_compile_flags --update --build --test --enable_onnx_tests --parallel
+
+      - name: Run ONNX Tests
+        shell: cmd
+        working-directory: ${{ runner.temp }}\build\RelWithDebInfo\RelWithDebInfo
+        run: |
+          .\onnx_test_runner -j 1 -e qnn -i "backend_path|%QNN_SDK_ROOT%\lib\x86_64-windows-msvc\QnnCpu.dll" ${{ github.workspace }}\cmake\external\onnx\onnx\backend\test\data\node
+
+      - name: Run float32 model tests
+        shell: cmd
+        working-directory: ${{ runner.temp }}\build\RelWithDebInfo\RelWithDebInfo
+        run: |
+          rem This step assumes the model data exists at C:\data\float32_models on the runner
+          if exist C:\data\float32_models (
+            .\onnx_test_runner -j 1 -e qnn -i "backend_path|%QNN_SDK_ROOT%\lib\x86_64-windows-msvc\QnnCpu.dll" C:\data\float32_models
+          ) else (
+            echo "Skipping float32 model tests: C:\data\float32_models not found."
+          )

.github/workflows/windows_webgpu.yml

@@ -127,22 +127,23 @@ jobs:
           }
           Remove-Item "${{ github.workspace }}\RelWithDebInfo" -Include "*.obj" -Recurse
 
-      - name: Run tests (onnxruntime_test_all) with verbose logging
+      - name: Run tests (onnxruntime_test_all, onnxruntime_provider_test) with verbose logging
         shell: pwsh
         run: |
           $env:ORT_UNIT_TEST_MAIN_LOG_LEVEL = "0"
-          .\onnxruntime_test_all.exe 2>.\onnxruntime_test_all_stderr.log
+          .\onnxruntime_test_all.exe 2> .\onnxruntime_test_stderr.log
+          .\onnxruntime_provider_test.exe 2>> .\onnxruntime_test_stderr.log
         working-directory: ${{ github.workspace }}\RelWithDebInfo\RelWithDebInfo
 
       - name: Check log file
         shell: cmd
         run: |
-          dir ${{ github.workspace }}\RelWithDebInfo\RelWithDebInfo\onnxruntime_test_all_stderr.log
+          dir ${{ github.workspace }}\RelWithDebInfo\RelWithDebInfo\onnxruntime_test_stderr.log
 
       - name: Validate shader keys
         uses: ./.github/actions/webgpu-validate-shader-key
         with:
-          log_file_path: ${{ github.workspace }}\RelWithDebInfo\RelWithDebInfo\onnxruntime_test_all_stderr.log
+          log_file_path: ${{ github.workspace }}\RelWithDebInfo\RelWithDebInfo\onnxruntime_test_stderr.log
 
       - name: Validate C# native delegates
         run: python tools\ValidateNativeDelegateAttributes.py

cmake/onnxruntime_mlas.cmake

@@ -109,6 +109,8 @@ function(setup_mlas_source_for_windows)
         ${MLAS_SRC_DIR}/eltwise_kernel_neon.cpp
         ${MLAS_SRC_DIR}/eltwise_kernel_neon_fp16.cpp
         ${MLAS_SRC_DIR}/sqnbitgemm_kernel_neon_int8_i8mm.cpp
+        ${MLAS_SRC_DIR}/sconv_kernel_neon.cpp
+        ${MLAS_SRC_DIR}/spool_kernel_neon.cpp
       )
 
       set(mlas_platform_preprocess_srcs
@@ -431,6 +433,8 @@ else()
           ${MLAS_SRC_DIR}/eltwise_kernel_neon.h
           ${MLAS_SRC_DIR}/eltwise_kernel_neon.cpp
           ${MLAS_SRC_DIR}/sqnbitgemm_kernel_neon_int8_i8mm.cpp
+          ${MLAS_SRC_DIR}/sconv_kernel_neon.cpp
+          ${MLAS_SRC_DIR}/spool_kernel_neon.cpp
         )
         if (onnxruntime_USE_KLEIDIAI)
           setup_kleidiai()

cmake/onnxruntime_test_pch.cmake

@@ -5,7 +5,10 @@ if(CMAKE_CXX_COMPILER_ID MATCHES "MSVC")
   target_precompile_headers(onnxruntime_test_all PRIVATE
     "${CMAKE_CURRENT_SOURCE_DIR}/test_pch.h"
   )
-  endif()
+  target_precompile_headers(onnxruntime_provider_test PRIVATE
+    "${CMAKE_CURRENT_SOURCE_DIR}/test_pch.h"
+  )
+endif()
 
 # Exclude certain files that might conflict with PCH
 set(PCH_EXCLUDE_FILES