Merge branch 'main' into dev/reenable_gluon_hangs

Author: HBN-MichalSzy

.github/workflows/build-test-reusable.yml

@@ -329,14 +329,6 @@ jobs:
           export PYTEST_MAX_PROCESSES=4
           ${{ env.TRITON_TEST_CMD }} --triton-kernels
 
-      # FIXME: make sure new tutorials are added to one of the groups (scaled_dot, rest, tutorial-faX)
-      - name: Select tutorials to run (scaled_dot)
-        if: matrix.suite == 'scaled_dot'
-        run: |
-          cat <<EOF | tee tutorials.txt
-          09-persistent-matmul
-          EOF
-
       - name: Select tutorials to run (rest)
         if: matrix.suite == 'rest'
         run: |
@@ -348,11 +340,12 @@ jobs:
           05-layer-norm
           07-extern-functions
           08-grouped-gemm
+          09-persistent-matmul
           10-experimental-block-pointer
           EOF
 
       - name: Run Tutorials
-        if: matrix.suite == 'scaled_dot' || matrix.suite == 'rest'
+        if: matrix.suite == 'rest'
         run: |
           ${{ env.TRITON_TEST_CMD }} --select-from-file tutorials.txt --tutorial
 

.github/workflows/build-test.yml

@@ -48,10 +48,16 @@ on:
     branches:
       - main
       - release/**
+    paths-ignore:
+      - '**/*.md'
+      - '.github/ISSUE_TEMPLATE/**'
   push:
     branches:
       - main
       - release/**
+    paths-ignore:
+      - '**/*.md'
+      - '.github/ISSUE_TEMPLATE/**'
 
 # Cancels in-progress PR runs when the PR is updated. Manual runs are never cancelled.
 concurrency:

.github/workflows/build-windows.yml

@@ -7,10 +7,16 @@ on:
     branches:
       - main
       - release/**
+    paths-ignore:
+      - '**/*.md'
+      - '.github/ISSUE_TEMPLATE/**'
   push:
     branches:
       - main
       - release/**
+    paths-ignore:
+      - '**/*.md'
+      - '.github/ISSUE_TEMPLATE/**'
 
 permissions: read-all
 

.github/workflows/inductor-tests.yml

@@ -52,6 +52,7 @@ env:
     inductor/test_select_algorithm.py
     inductor/test_max_autotune.py
     inductor/test_compile_subprocess.py
+    inductor/test_analysis.py
 
 jobs:
   compute-params:

.github/workflows/integration-tests-amd.yml

@@ -13,31 +13,33 @@ jobs:
   integration-tests-amd:
     runs-on: ${{ matrix.runner }}
     timeout-minutes: 45
-    continue-on-error: ${{ matrix.runner[1] == 'gfx90a' }}
+    continue-on-error: ${{ matrix.runner[1] == 'gfx90a' || matrix.runner[0] == 'amd-gfx950' }}
     strategy:
       matrix:
         runner: ${{ fromJson(inputs.matrix) }}
         include:
-          - image: rocm/pytorch:rocm6.2.2_ubuntu22.04_py3.10_pytorch_2.5.1_asan
+          - image: rocm/pytorch:rocm7.0_ubuntu22.04_py3.10_pytorch_release_2.8.0
             runner: ["self-hosted", "gfx90a"]
             # Cache save/restore is on the host machine at directory /home/runner/.triton, while in the docker
             # container expect it at /github/home/.triton. So map here to make sure visible in docker.
             options: >-
               --device=/dev/kfd --device=/dev/dri --security-opt seccomp=unconfined --group-add video --user root
               --volume /home/runner/.triton:/github/home/.triton
-          - image: rocm/pytorch:rocm6.2.2_ubuntu22.04_py3.10_pytorch_2.5.1_asan
+          - image: rocm/pytorch:rocm7.0_ubuntu22.04_py3.10_pytorch_release_2.8.0
             runner: ["amd-gfx942"]
             # We add --env-file to pull in HIP_VISIBLE_DEVICES and ROCR_VISIBLE_DEVICES definition for GPU isolation.
             options: >-
               --device=/dev/kfd --device=/dev/dri --security-opt seccomp=unconfined --group-add video --user root
               --env-file /etc/podinfo/gha-gpu-isolation-settings
               --volume /home/runner/.triton:/github/home/.triton
-          - image: rocm/7.0-preview:rocm7.0_preview_ubuntu22.04_llama2_70b_training_mlperf_mi35X_prealpha
+          - image: rocm/pytorch:rocm7.0_ubuntu22.04_py3.10_pytorch_release_2.8.0
             runner: ["amd-gfx950"]
+            # We add --env-file to pull in HIP_VISIBLE_DEVICES and ROCR_VISIBLE_DEVICES definition for GPU isolation.
             options: >-
               --device=/dev/kfd --device=/dev/dri --security-opt seccomp=unconfined --group-add video --user root
               --env-file /etc/podinfo/gha-gpu-isolation-settings
               --volume /home/runner/.triton:/github/home/.triton
+              --volume /triton-data:/triton-data
     env:
       RUNNER_TYPE: ${{ matrix.runner[1] }}
       TRITON_BUILD_WITH_CCACHE: "true"
@@ -83,14 +85,16 @@ jobs:
             ~/.triton/nvidia
             ~/.triton/json
           key: ${{ runner.os }}-${{ runner.arch }}-llvm-${{ steps.cache-key.outputs.llvm }}-nvidia-${{ steps.cache-key.outputs.nvidia }}-json-${{ steps.cache-key.outputs.json }}
+      - name: Install dependencies
+        run: apt-get install -y clang lld ccache
       - name: Inspect cache directories
         run: |
           mkdir -p ~/.triton
           du -h -d 1 ~/.triton
 
           mkdir -p ~/.ccache
           du -h -d 1 ~/.ccache
-      - name: Update compiler to clang
+      - name: Update compiler to Clang
         run: |
           export CC=/usr/bin/clang
           export CXX=/usr/bin/clang++
@@ -100,19 +104,19 @@ jobs:
           echo "PATH is '$PATH'"
           pip uninstall -y triton pytorch-triton-rocm
 
-          if [ "${{ matrix.runner[0] }}" != "amd-gfx950" ]; then
-            ccache --zero-stats
+          ccache --zero-stats
+          if [ "${{ matrix.runner[0] }}" = "amd-gfx950" ]; then
+            pip install --cache-dir /triton-data/pip-cache -r python/requirements.txt
+            pip install --cache-dir /triton-data/pip-cache -r python/test-requirements.txt
           fi
-
           make dev-install
-      - name: CCache Stats
-        if: ${{ matrix.runner[0] != 'amd-gfx950' }}
+      - name: Print ccache stats
         run: ccache --print-stats
       - name: Run lit tests
         run: make test-lit
       - name: Run C++ unittests
         run: make test-cpp
-      - name: Run python tests on AMD
+      - name: Run Python tests on AMD
         run: |
           INSTRUMENTATION_LIB_DIR="${GITHUB_WORKSPACE}/python/triton/instrumentation"
           if [ ! -d "${INSTRUMENTATION_LIB_DIR}" ]; then

.github/workflows/spirvrunner-test.yml

@@ -6,9 +6,15 @@ on:
   pull_request:
     branches:
       - main
+    paths-ignore:
+      - '**/*.md'
+      - '.github/ISSUE_TEMPLATE/**'
   push:
     branches:
       - main
+    paths-ignore:
+      - '**/*.md'
+      - '.github/ISSUE_TEMPLATE/**'
 
 permissions: read-all
 

.github/workflows/third-party-benchmarks.yml

@@ -223,6 +223,18 @@ jobs:
             --max-new-tokens $MAX_NEW_TOKENS \
             --batch-size $BATCH_SIZE
 
+      - name: Run launch microbenchmark tests
+        if: ${{ steps.install.outcome == 'success' && !cancelled() && (inputs.benchmarks == '' || contains(fromJson(inputs.benchmarks || '[]'), 'launch_micro_benchmarks')) && !contains(fromJson(inputs.skip_benchmarks || '[]'), 'launch_micro_benchmarks') }}
+        run: |
+          source scripts/capture-hw-details.sh
+          python python/test/microbenchmark/launch_overhead.py --reports $REPORTS
+
+          python benchmarks/third_party/vllm/transform_results.py $REPORTS/launch_overhead_results.csv $REPORTS/launch_overhead-report.csv \
+            --tag $TAG \
+            --bgroup overhead \
+            --benchmark launch-overhead \
+            --param_cols="input_type"
+
       - name: Upload benchmark reports
         if: ${{ steps.install.outcome == 'success' && !cancelled() }}
         uses: actions/upload-artifact@v5

.github/workflows/triton-benchmarks.yml

@@ -65,6 +65,9 @@ env:
   VERIFY: ${{ (github.event_name == 'pull_request' || github.event_name == 'schedule' || inputs.verify) && '1' || '0' }}
   TAG: ${{ inputs.tag || (github.event_name == 'pull_request' && format('pr-{0}', github.event.number)) || (github.event_name == 'schedule' && 'ci') || 'test' }}
   N_RUNS: ${{ inputs.n_runs || '1' }}
+  # FIXME: Enable Level Zero v2 loader once it's stable.
+  # https://github.com/intel/intel-xpu-backend-for-triton/issues/5572
+  UR_LOADER_USE_LEVEL_ZERO_V2: "0"
 
 jobs:
   build:

.github/workflows/try-latest-pytorch.yml

@@ -96,6 +96,7 @@ jobs:
         inductor/test_select_algorithm.py
         inductor/test_max_autotune.py
         inductor/test_compile_subprocess.py
+        inductor/test_analysis.py
       runner_label: ${{ inputs.runner_label }}
       python_version: "3.10"
 

.gitignore

@@ -51,6 +51,9 @@ pytest.ini
 # Instrumentation
 python/triton/instrumentation
 
+# MLIR Plugin
+python/triton/plugins
+
 # Python caches
 __pycache__/
 *.py[cod]