Merge remote-tracking branch 'origin/sycl' into upd-is-comp

Author: KornevNikita

.github/CODEOWNERS

@@ -36,10 +36,8 @@ sycl/doc/extensions/ @intel/dpcpp-specification-reviewers
 
 # Unified Runtime
 unified-runtime/ @intel/unified-runtime-reviewers
-# TODO: Use specific UR Level Zero adapter team
-unified-runtime/*/adapters/level_zero/ @intel/unified-runtime-reviewers
-# TODO: Use specific UR OpenCL adapter team
-unified-runtime/*/adapters/opencl/ @intel/unified-runtime-reviewers
+unified-runtime/*/adapters/level_zero/ @intel/unified-runtime-reviewers-level-zero
+unified-runtime/*/adapters/opencl/ @intel/unified-runtime-reviewers-opencl
 unified-runtime/*/adapters/cuda/  @intel/llvm-reviewers-cuda
 unified-runtime/*/adapters/hip/  @intel/llvm-reviewers-cuda
 unified-runtime/*/adapters/native_cpu/ @intel/dpcpp-nativecpu-reviewers
@@ -75,6 +73,7 @@ llvm/include/llvm/SYCLLowerIR/SYCLCreateNVVMAnnotations.h @intel/llvm-reviewers-
 llvm/lib/SYCLLowerIR/SYCLCreateNVVMAnnotations.cpp @intel/llvm-reviewers-cuda
 llvm/lib/Target/NVPTX @intel/llvm-reviewers-cuda
 llvm/lib/Target/AMDGPU @intel/llvm-reviewers-cuda
+unified-runtime/source/common/cuda-hip @intel/llvm-reviewers-cuda
 
 # XPTI instrumentation utilities
 xpti/ @intel/llvm-reviewers-runtime

.github/workflows/sycl-detect-changes.yml

@@ -82,6 +82,8 @@ jobs:
             ur:
               - 'unified-runtime/**'
               - .github/workflows/ur-*
+            ur_cuda_adapter:
+              - 'unified-runtime/source/adapters/cuda/**'
 
       - name: Set output
         id: result
@@ -92,7 +94,7 @@ jobs:
               return '${{ steps.changes.outputs.changes }}';
             }
             // Treat everything as changed for huge PRs.
-            return ["llvm", "llvm_spirv", "clang", "sycl_jit", "xptifw", "libclc", "sycl", "ci", "esimd", "ur"];
+            return ["llvm", "llvm_spirv", "clang", "sycl_jit", "xptifw", "libclc", "sycl", "ci", "esimd", "ur", "ur_cuda_adapter"];
 
       - run: echo '${{ steps.result.outputs.result }}'
 

.github/workflows/sycl-linux-build.yml

@@ -46,7 +46,10 @@ on:
         default: 3
       e2e_binaries_artifact:
         type: string
-        required: False
+        required: false
+      pack_release:
+        type: string
+        required: false
 
     outputs:
       build_conclusion:
@@ -100,6 +103,10 @@ on:
         options:
           - 3
 
+      pack_release:
+        type: string
+        required: false
+
 permissions: read-all
 
 jobs:
@@ -215,11 +222,26 @@ jobs:
         # TODO consider moving this to Dockerfile.
         export LD_LIBRARY_PATH=/usr/local/cuda/compat/:/usr/local/cuda/lib64:$LD_LIBRARY_PATH
         LIT_OPTS="--allow-empty-runs" LIT_FILTER="e2e_test_requirements" cmake --build $GITHUB_WORKSPACE/build --target check-sycl
-    - name: Install
+    - name: Install sycl-toolchain
       if: ${{ always() && !cancelled() && steps.build.conclusion == 'success' }}
-      # TODO replace utility installation with a single CMake target
       run: |
         cmake --build $GITHUB_WORKSPACE/build --target deploy-sycl-toolchain
+
+    - name: Pack toolchain release
+      if: ${{ always() && !cancelled() && steps.build.conclusion == 'success' && inputs.pack_release == 'true' }}
+      run: tar -I '${{ steps.artifact_info.outputs.COMPRESS }}' -cf ${{ steps.artifact_info.outputs.ARCHIVE_NAME }} -C $GITHUB_WORKSPACE/build/install .
+    - name: Upload toolchain release
+      if: ${{ always() && !cancelled() && steps.build.conclusion == 'success' && inputs.pack_release == 'true' }}
+      uses: actions/upload-artifact@v4
+      with:
+        name: sycl_linux_release
+        path: ${{ steps.artifact_info.outputs.ARCHIVE_NAME }}
+        retention-days: ${{ inputs.retention-days }}
+
+    - name: Install utilities
+      if: ${{ always() && !cancelled() && steps.build.conclusion == 'success' }}
+      # TODO replace utility installation with a single CMake target
+      run: |
         cmake --build $GITHUB_WORKSPACE/build --target utils/FileCheck/install
         cmake --build $GITHUB_WORKSPACE/build --target utils/count/install
         cmake --build $GITHUB_WORKSPACE/build --target utils/not/install

.github/workflows/sycl-linux-precommit.yml

@@ -22,8 +22,14 @@ on:
     - '.github/workflows/sycl-macos-*.yml'
     - '.github/workflows/sycl-nightly.yml'
     - '.github/workflows/sycl-rel-nightly.yml'
+    - '.github/workflows/trivy.yml'
     - 'devops/containers/**'
     - 'devops/actions/build_container/**'
+    - 'unified-runtime/examples/**'
+    - 'unified-runtime/scripts/**'
+    - 'unified-runtime/test/**'
+    - 'unified-runtime/third_party/**'
+    - 'unified-runtime/tools/**'
 
 concurrency:
   #  Cancel a currently running workflow from the same PR, branch or tag.
@@ -52,6 +58,22 @@ jobs:
       changes: ${{ needs.detect_changes.outputs.filters }}
       e2e_binaries_artifact: sycl_e2e_bin_default
 
+  # If a PR changes CUDA adapter, run the build on Ubuntu 22.04 as well.
+  # Ubuntu 22.04 container has CUDA 12.1 installed while Ubuntu 24.0 image
+  # has CUDA 12.6.1 installed.
+  # The idea is to ensure that the code works with both CUDA versions.
+  build_ubuntu2204:
+    needs: [detect_changes]
+    if: always() && !cancelled() && contains(needs.detect_changes.outputs.filters, 'ur_cuda_adapter')
+    uses: ./.github/workflows/sycl-linux-build.yml
+    with:
+      build_ref: ${{ github.sha }}
+      build_cache_root: "/__w/"
+      build_artifact_suffix: "default"
+      build_cache_suffix: "default"
+      build_image: "ghcr.io/intel/llvm/ubuntu2204_build:latest"
+      changes: ${{ needs.detect_changes.outputs.filters }}
+
   run_prebuilt_e2e_tests:
     needs: [build, detect_changes]
     if: ${{ always() && !cancelled() && needs.build.outputs.build_conclusion == 'success' }}

.github/workflows/sycl-rel-nightly.yml

@@ -63,34 +63,35 @@ jobs:
             runner: '["Linux", "amdgpu"]'
             image_options: -u 1001 --device=/dev/dri --device=/dev/kfd
             target_devices: hip:gpu
-            tests_selector: e2e
 
-          - name: Intel L0 GPU
+          - name: Intel L0 Gen12 GPU
             runner: '["Linux", "gen12"]'
             image_options: -u 1001 --device=/dev/dri -v /dev/dri/by-path:/dev/dri/by-path --privileged --cap-add SYS_ADMIN
             target_devices: level_zero:gpu
-            tests_selector: e2e
             extra_lit_opts: --param gpu-intel-gen12=True
 
-          - name: Intel OCL GPU
+          - name: Intel L0 Battlemage GPU
+            runner: '["Linux", "bmg"]'
+            image_options: -u 1001 --device=/dev/dri -v /dev/dri/by-path:/dev/dri/by-path --privileged --cap-add SYS_ADMIN
+            target_devices: level_zero:gpu
+
+          - name: Intel OCL Gen12 GPU
             runner: '["Linux", "gen12"]'
             image_options: -u 1001 --device=/dev/dri -v /dev/dri/by-path:/dev/dri/by-path --privileged --cap-add SYS_ADMIN
             target_devices: opencl:gpu
-            tests_selector: e2e
             extra_lit_opts: --param gpu-intel-gen12=True
 
           - name: Intel OCL CPU
             runner: '["Linux", "gen12"]'
             image_options: -u 1001 --privileged --cap-add SYS_ADMIN
             target_devices: opencl:cpu
-            tests_selector: e2e
     uses: ./.github/workflows/sycl-linux-run-tests.yml
     with:
       name: ${{ matrix.name }}
       runner: ${{ matrix.runner }}
       image_options: ${{ matrix.image_options }}
       target_devices: ${{ matrix.target_devices }}
-      tests_selector: ${{ matrix.tests_selector }}
+      tests_selector: e2e
       extra_lit_opts: ${{ matrix.extra_lit_opts }}
       repo_ref: ${{ inputs.testing_branch || 'sycl-rel-6_1_0'  }}
       devops_ref: sycl
@@ -117,12 +118,22 @@ jobs:
       always()
       && !cancelled()
       && needs.build-win.outputs.build_conclusion == 'success'
+    strategy:
+      fail-fast: false
+      matrix:
+        include:
+          - name: Intel GEN12 Graphics with Level Zero
+            runner: '["Windows","gen12"]'
+            # 6_1_0 doesn't support the auto-detection, keep it until 6_2_0.
+            extra_lit_opts: --param gpu-intel-gen12=True
+          - name: Intel Battlemage Graphics with Level Zero
+            runner: '["Windows","bmg"]'
     uses: ./.github/workflows/sycl-windows-run-tests.yml
     with:
-      name: Intel GEN12 Graphics with Level Zero
-      runner: '["Windows","gen12"]'
+      name: ${{ matrix.name }}
+      runner: ${{ matrix.runner }}
       sycl_toolchain_archive: ${{ needs.build-win.outputs.artifact_archive_name }}
-      extra_lit_opts: --param gpu-intel-gen12=True
+      extra_lit_opts: ${{ matrix.extra_lit_opts }}
       repo_ref: ${{ inputs.testing_branch || 'sycl-rel-6_1_0'  }}
       devops_ref: sycl
 
@@ -203,7 +214,7 @@ jobs:
       image_options: ${{ matrix.image_options }}
       target_devices: ${{ matrix.target_devices }}
       tests_selector: cts
-      repo_ref: ${{ github.sha }}
+      repo_ref: ${{ inputs.testing_branch || 'sycl-rel-6_1_0'  }}
       devops_ref: sycl
       sycl_toolchain_artifact: sycl_linux_default
       sycl_toolchain_archive: ${{ needs.ubuntu2204_build.outputs.artifact_archive_name }}

.github/workflows/sycl-windows-precommit.yml

@@ -21,8 +21,14 @@ on:
     - '.github/workflows/sycl-macos-*.yml'
     - '.github/workflows/sycl-nightly.yml'
     - '.github/workflows/sycl-rel-nightly.yml'
+    - '.github/workflows/trivy.yml'
     - 'devops/containers/**'
     - 'devops/actions/build_container/**'
+    - 'unified-runtime/examples/**'
+    - 'unified-runtime/scripts/**'
+    - 'unified-runtime/test/**'
+    - 'unified-runtime/third_party/**'
+    - 'unified-runtime/tools/**'
 
 permissions: read-all
 

.github/workflows/sycl-windows-run-tests.yml

@@ -117,12 +117,14 @@ jobs:
         echo "C:\Program Files\Git\usr\bin" | Out-File -FilePath $env:GITHUB_PATH -Encoding utf8 -Append
     - name: Register cleanup after job is finished
       uses: ./devops/actions/cleanup
-    - uses: ./devops/actions/cached_checkout
+    - uses: actions/checkout@v4
       if: inputs.tests_selector == 'e2e'
       with:
         path: llvm
         ref: ${{ inputs.repo_ref || github.sha }}
-        cache_path: "D:\\\\github\\\\_work\\\\repo_cache\\\\"
+        sparse-checkout: |
+          llvm/utils/lit
+          sycl/test-e2e          
     - name: Download compiler toolchain
       uses: actions/download-artifact@v4
       with:

.github/workflows/trivy.yml

@@ -30,6 +30,7 @@ jobs:
         run: ./bin/trivy config --format json --output trivy-report.json --ignorefile=devops/.trivyignore.yaml devops/containers --exit-code 1
 
       - name: Upload report artifact
+        if: always() && !cancelled()
         uses: actions/upload-artifact@v4
         with:
           name: trivy-report

clang/lib/Driver/ToolChains/Clang.cpp

@@ -5885,8 +5885,16 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA,
         CmdArgs.push_back("-ffine-grained-bitfield-accesses");
     }
 
-    if (!Args.hasFlag(options::OPT_fsycl_unnamed_lambda,
-                      options::OPT_fno_sycl_unnamed_lambda, true))
+    // '-fsycl-unnamed-lambda' is not supported with '-fsycl-host-compiler'
+    if (Args.hasArg(options::OPT_fsycl_host_compiler_EQ)) {
+      if (Args.hasFlag(options::OPT_fsycl_unnamed_lambda,
+                       options::OPT_fno_sycl_unnamed_lambda, false))
+        D.Diag(diag::err_drv_cannot_mix_options) << "-fsycl-host-compiler"
+                                                 << "-fsycl-unnamed-lambda";
+      else // '-fsycl-host-compiler' implies '-fno-sycl-unnamed-lambda'
+        CmdArgs.push_back("-fno-sycl-unnamed-lambda");
+    } else if (!Args.hasFlag(options::OPT_fsycl_unnamed_lambda,
+                             options::OPT_fno_sycl_unnamed_lambda, true))
       CmdArgs.push_back("-fno-sycl-unnamed-lambda");
 
     if (!Args.hasFlag(options::OPT_fsycl_esimd_force_stateless_mem,

clang/test/CodeGenSYCL/sub-group-size.cpp

@@ -15,15 +15,15 @@ using namespace sycl;
 // ALL-DAG: define {{.*}}spir_func void @{{.*}}external_10{{.*}}() #{{[0-9]+}}{{.*}} !intel_reqd_sub_group_size ![[TEN:[0-9]+]]
 
 SYCL_EXTERNAL void external_default_behavior() {}
-// NONE-DAG: define {{.*}}spir_func void @{{.*}}external_default_behavior{{.*}}() #{{[0-9]+}} !srcloc !{{[0-9]+}} !sycl_fixed_targets !{{[0-9]+}} {
+// NONE-DAG: define {{.*}}spir_func void @{{.*}}external_default_behavior{{.*}}() #{{[0-9]+}} !sycl_fixed_targets !{{[0-9]+}} {
 // PRIM_DEF-DAG: define {{.*}}spir_func void @{{.*}}external_default_behavior{{.*}}() #{{[0-9]+}}{{.*}} !intel_reqd_sub_group_size ![[PRIMARY]]
 // TEN_DEF-DAG: define {{.*}}spir_func void @{{.*}}external_default_behavior{{.*}}() #{{[0-9]+}}{{.*}} !intel_reqd_sub_group_size ![[TEN]]
 
 void default_behavior() {
   kernel_single_task<class Kernel1>([]() {
   });
 }
-// NONE-DAG: define {{.*}}spir_kernel void @{{.*}}Kernel1() #{{[0-9]+}} !srcloc !{{[0-9]+}} !kernel_arg_buffer_location !{{[0-9]+}} !sycl_fixed_targets !{{[0-9]+}} {
+// NONE-DAG: define {{.*}}spir_kernel void @{{.*}}Kernel1() #{{[0-9]+}} !kernel_arg_buffer_location !{{[0-9]+}} !sycl_fixed_targets !{{[0-9]+}} {
 // PRIM_DEF-DAG: define {{.*}}spir_kernel void @{{.*}}Kernel1() #{{[0-9]+}}{{.*}} !intel_reqd_sub_group_size ![[PRIMARY]]
 // TEN_DEF-DAG: define {{.*}}spir_kernel void @{{.*}}Kernel1() #{{[0-9]+}}{{.*}} !intel_reqd_sub_group_size ![[TEN]]