Merge branch 'main' into feature/2246-e2e-accuracy-subset

Author: kwasd

.github/models/accuracy/huggingface.txt

@@ -0,0 +1,10 @@
+BlenderbotSmallForCausalLM
+CamemBert
+DistilBertForMaskedLM
+ElectraForCausalLM
+GPT2ForSequenceClassification
+MobileBertForMaskedLM
+MobileBertForQuestionAnswering
+Speech2Text2ForCausalLM
+T5ForConditionalGeneration
+XLNetLMHeadModel

.github/models/accuracy/timm_models.txt

@@ -0,0 +1,6 @@
+beit_base_patch16_224
+fbnetv3_b
+gmixer_24_224
+mobilenetv3_large_100
+nfnet_l0
+tnt_s_patch16_224

.github/models/accuracy/torchbench.txt

@@ -0,0 +1,4 @@
+resnet50
+hf_T5
+densenet121
+mobilenet_v2

.github/pins/pytorch-upstream.txt

@@ -1 +1 @@
-0a2685160140656e3e53818611dd2c65c4397be5
+8321eec009c8c79145ebccd51fdfc336e5f8b848

.github/workflows/auto-update-translator-cid.yml

@@ -24,6 +24,7 @@ jobs:
         uses: actions/checkout@v4
         with:
           fetch-depth: 0
+          token: ${{ secrets.BOT_TOKEN }}
 
       # search the opened PR that has branch "bot/update_translator_cid", no new PR as long as the target PR is open
       - name: Search the opened PR that has branch ${{ env.PR_BRANCH }}
@@ -98,8 +99,8 @@ jobs:
         env:
           GITHUB_TOKEN: ${{ secrets.BOT_TOKEN }}
         run: |
-          git config --global user.name 'github-actions[bot]'
-          git config --global user.email 'github-actions[bot]@users.noreply.github.com'
+          git config --global user.name 'glados-intel'
+          git config --global user.email 'glados@intel.com'
 
           if git show-ref --verify --quiet "refs/remotes/origin/${PR_BRANCH}"; then
             echo "${PR_BRANCH} already exist, deleting..."

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

@@ -159,7 +159,6 @@ jobs:
           echo "TRANSFORMERS_VERSION=$TRANSFORMERS_VERSION" | tee -a $GITHUB_ENV
 
       - name: Install transformers
-        if: ${{ inputs.python_version != '3.12' }}
         uses: ./.github/actions/install-dependency
         with:
           package: transformers
@@ -168,7 +167,6 @@ jobs:
           try-tag-prefix: v
 
       - name: Run E2E test
-        if: ${{ inputs.python_version != '3.12' }}
         run: |
           # Set WORKSPACE for inductor_xpu_test.sh to make sure it creates "inductor_log" outside of pytorch cloned directory
           export WORKSPACE=$GITHUB_WORKSPACE

.github/workflows/e2e-reusable.yml

@@ -222,16 +222,15 @@ jobs:
           cd pytorch
 
           # if "only_one_model" is set, then test this model
-          # if "models" == "subset", then test the models from .github/models/performance/{suite}.txt
+          # if "models" == "subset", then test the models from .github/models/{accuracy,performance}/{suite}.txt
           # otherwise test all models.
 
           if [[ "${{ inputs.only_one_model }}" ]]; then
             bash -e $GITHUB_WORKSPACE/scripts/inductor_xpu_test.sh ${{ inputs.suite }} ${{ inputs.dtype }} ${{ inputs.mode }} ${{ inputs.test_mode }} xpu 0 static 1 0 ${{ inputs.only_one_model }}
           elif [[ "${{ inputs.models }}" == "subset" ]]; then
             while read model; do
               bash -e $GITHUB_WORKSPACE/scripts/inductor_xpu_test.sh ${{ inputs.suite }} ${{ inputs.dtype }} ${{ inputs.mode }} ${{ inputs.test_mode }} xpu 0 static 1 0 $model
-              grep ,$model, $WORKSPACE/inductor_log/*/*/*.csv | grep -q ,pass, || ${{ inputs.test_all_subset_models }}
-            done < $GITHUB_WORKSPACE/.github/models/performance/${{ inputs.suite }}.txt
+            done < $GITHUB_WORKSPACE/.github/models/${{ inputs.test_mode }}/${{ inputs.suite }}.txt
           else
             bash -e $GITHUB_WORKSPACE/scripts/inductor_xpu_test.sh ${{ inputs.suite }} ${{ inputs.dtype }} ${{ inputs.mode }} ${{ inputs.test_mode }} xpu 0 static 1 0
           fi

.github/workflows/integration-tests.yml

@@ -226,7 +226,7 @@ jobs:
       - name: Install pip dependencies
         run: |
           python3 -m pip install --upgrade pip
-          python3 -m pip install wheel cmake==3.24 ninja pytest-xdist lit
+          python3 -m pip install cython setuptools wheel cmake==3.24 ninja pytest-forked pytest-xdist lit
       - name: Install Triton
         env:
           TRITON_BUILD_WITH_CCACHE: "true"
@@ -245,18 +245,19 @@ jobs:
           lit -v "${LIT_TEST_DIR}"
       - name: Run python tests on CUDA
         run: |
-          SHARED_LIB_DIR="${GITHUB_WORKSPACE}/python/build/$(ls python/build | grep -i lib)/triton/_C"
-          if [ ! -d "${SHARED_LIB_DIR}" ]; then
-            echo "Coult not find '${SHARED_LIB_DIR}'" ; exit -1
+          INSTRUMENTATION_LIB_DIR="${GITHUB_WORKSPACE}/python/build/$(ls python/build | grep -i lib)/triton/instrumentation"
+          if [ ! -d "${INSTRUMENTATION_LIB_DIR}" ]; then
+            echo "Coult not find '${INSTRUMENTATION_LIB_DIR}'" ; exit -1
           fi
           cd python/test/unit
-          python3 -m pytest -s -n 8 --ignore=hopper/test_flashattention.py --ignore=language/test_line_info.py --ignore=language/test_subprocess.py
+          python3 -m pytest -s -n 8 --ignore=hopper/test_flashattention.py --ignore=language/test_line_info.py --ignore=language/test_subprocess.py --ignore=test_debug.py
           python3 -m pytest -s -n 8 language/test_subprocess.py
+          python3 -m pytest -s -n 8 test_debug.py --forked
           # Run test_line_info.py separately with TRITON_DISABLE_LINE_INFO=0
           TRITON_DISABLE_LINE_INFO=0 python3 -m pytest -s language/test_line_info.py
           # Run hopper/test_flashattention.py separately to avoid out of gpu memory
           python3 -m pytest -s hopper/test_flashattention.py
-          TRITON_ALWAYS_COMPILE=1 TRITON_DISABLE_LINE_INFO=0 LLVM_PASS_PLUGIN_PATH=${SHARED_LIB_DIR}/libGPUHello.so \
+          TRITON_ALWAYS_COMPILE=1 TRITON_DISABLE_LINE_INFO=0 LLVM_PASS_PLUGIN_PATH=${INSTRUMENTATION_LIB_DIR}/libGPUInstrumentationTestLib.so \
           python3 -m pytest --capture=tee-sys -rfs -vvv instrumentation/test_gpuhello.py
       - name: Run interpreter tests
         if: ${{ matrix.runner[0] == 'h100-runner-set' }}
@@ -400,15 +401,18 @@ jobs:
           lit -v "${LIT_TEST_DIR}"
       - name: Run python tests on HIP
         run: |
-          SHARED_LIB_DIR="${GITHUB_WORKSPACE}/python/triton/_C"
-          if [ ! -d "${SHARED_LIB_DIR}" ]; then
-            echo "Coult not find '${SHARED_LIB_DIR}'" ; exit -1
+          INSTRUMENTATION_LIB_DIR="${GITHUB_WORKSPACE}/python/triton/instrumentation"
+          if [ ! -d "${INSTRUMENTATION_LIB_DIR}" ]; then
+            echo "Coult not find '${INSTRUMENTATION_LIB_DIR}'" ; exit -1
           fi
           pytest --capture=tee-sys -rfs python/tutorials/06-fused-attention.py
           cd python/test/unit
           pytest --capture=tee-sys -rfs -n 16 language runtime \
-                 --ignore=language/test_line_info.py
-          TRITON_ALWAYS_COMPILE=1 TRITON_DISABLE_LINE_INFO=0 LLVM_PASS_PLUGIN_PATH=${SHARED_LIB_DIR}/libGPUHello.so \
+                 --ignore=language/test_line_info.py \
+                 --ignore=test_debug.py
+          # TODO: uncomment
+          # pytest --capture=tee-sys -rfs test_debug.py
+          TRITON_ALWAYS_COMPILE=1 TRITON_DISABLE_LINE_INFO=0 LLVM_PASS_PLUGIN_PATH=${INSTRUMENTATION_LIB_DIR}/libGPUInstrumentationTestLib.so \
           pytest --capture=tee-sys -rfs -vvv instrumentation/test_gpuhello.py
 
           # Run test_line_info.py separately with TRITON_DISABLE_LINE_INFO=0

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

@@ -256,7 +256,7 @@ jobs:
       - name: Install pip dependencies
         run: |
           python3 -m pip install --upgrade pip
-          python3 -m pip install wheel cmake==3.24 ninja pytest-xdist lit
+          python3 -m pip install cython setuptools wheel cmake==3.24 ninja pytest-forked pytest-xdist lit
 
       - name: Install Triton
         env:
@@ -279,18 +279,19 @@ jobs:
 
       - name: Run python tests on CUDA
         run: |
-          SHARED_LIB_DIR="${GITHUB_WORKSPACE}/python/build/$(ls python/build | grep -i lib)/triton/_C"
-          if [ ! -d "${SHARED_LIB_DIR}" ]; then
-            echo "Coult not find '${SHARED_LIB_DIR}'" ; exit -1
+          INSTRUMENTATION_LIB_DIR="${GITHUB_WORKSPACE}/python/build/$(ls python/build | grep -i lib)/triton/instrumentation"
+          if [ ! -d "${INSTRUMENTATION_LIB_DIR}" ]; then
+            echo "Coult not find '${INSTRUMENTATION_LIB_DIR}'" ; exit -1
           fi
           cd python/test/unit
-          python3 -m pytest -s -n 8 --ignore=hopper/test_flashattention.py --ignore=language/test_line_info.py --ignore=language/test_subprocess.py
+          python3 -m pytest -s -n 8 --ignore=hopper/test_flashattention.py --ignore=language/test_line_info.py --ignore=language/test_subprocess.py --ignore=test_debug.py
           python3 -m pytest -s -n 8 language/test_subprocess.py
+          python3 -m pytest -s -n 8 test_debug.py --forked
           # Run test_line_info.py separately with TRITON_DISABLE_LINE_INFO=0
           TRITON_DISABLE_LINE_INFO=0 python3 -m pytest -s language/test_line_info.py
           # Run hopper/test_flashattention.py separately to avoid out of gpu memory
           python3 -m pytest -s hopper/test_flashattention.py
-          TRITON_ALWAYS_COMPILE=1 TRITON_DISABLE_LINE_INFO=0 LLVM_PASS_PLUGIN_PATH=${SHARED_LIB_DIR}/libGPUHello.so \
+          TRITON_ALWAYS_COMPILE=1 TRITON_DISABLE_LINE_INFO=0 LLVM_PASS_PLUGIN_PATH=${INSTRUMENTATION_LIB_DIR}/libGPUInstrumentationTestLib.so \
           python3 -m pytest --capture=tee-sys -rfs -vvv instrumentation/test_gpuhello.py
 
       - name: Run interpreter tests
@@ -396,15 +397,18 @@ jobs:
 
       - name: Run python tests on HIP
         run: |
-          SHARED_LIB_DIR="${GITHUB_WORKSPACE}/python/triton/_C"
-          if [ ! -d "${SHARED_LIB_DIR}" ]; then
-            echo "Coult not find '${SHARED_LIB_DIR}'" ; exit -1
+          INSTRUMENTATION_LIB_DIR="${GITHUB_WORKSPACE}/python/triton/instrumentation"
+          if [ ! -d "${INSTRUMENTATION_LIB_DIR}" ]; then
+            echo "Coult not find '${INSTRUMENTATION_LIB_DIR}'" ; exit -1
           fi
           pytest --capture=tee-sys -rfs python/tutorials/06-fused-attention.py
           cd python/test/unit
           pytest --capture=tee-sys -rfs -n 16 language runtime \
-                 --ignore=language/test_line_info.py
-          TRITON_ALWAYS_COMPILE=1 TRITON_DISABLE_LINE_INFO=0 LLVM_PASS_PLUGIN_PATH=${SHARED_LIB_DIR}/libGPUHello.so \
+                 --ignore=language/test_line_info.py \
+                 --ignore=test_debug.py
+          # TODO: uncomment
+          # pytest --capture=tee-sys -rfs test_debug.py
+          TRITON_ALWAYS_COMPILE=1 TRITON_DISABLE_LINE_INFO=0 LLVM_PASS_PLUGIN_PATH=${INSTRUMENTATION_LIB_DIR}/libGPUInstrumentationTestLib.so \
           pytest --capture=tee-sys -rfs -vvv instrumentation/test_gpuhello.py
 
           # Run test_line_info.py separately with TRITON_DISABLE_LINE_INFO=0

.github/workflows/llvm-build.yml

@@ -107,7 +107,6 @@ jobs:
         -DLLVM_INSTALL_UTILS=ON
         -DLLVM_TARGETS_TO_BUILD="host;NVPTX;AMDGPU"
         -DLLVM_ENABLE_TERMINFO=OFF
-        -DLLVM_ABI_BREAKING_CHECKS=FORCE_OFF
         llvm-project/llvm
 
         ninja -C llvm-project/build check-mlir install
@@ -131,7 +130,6 @@ jobs:
         -DLLVM_INSTALL_UTILS=ON
         -DLLVM_TARGETS_TO_BUILD="host;NVPTX;AMDGPU"
         -DLLVM_ENABLE_TERMINFO=OFF
-        -DLLVM_ABI_BREAKING_CHECKS=FORCE_OFF
         llvm-project/llvm
 
         ninja -C llvm-project/build check-mlir install