add cuda test

Author: jinyan-li1

.github/workflows/pr-vllm.yml

@@ -153,3 +153,79 @@ jobs:
           docker rm -f ${CONTAINER_ID} || true
           docker image prune -a --force --filter "until=24h"
           docker system df
+
+  cuda-test:
+    needs: [build-image]
+    if: needs.build-image.result == 'success'
+    runs-on:
+      - codebuild-runner-${{ github.run_id }}-${{ github.run_attempt }}
+        fleet:x86-g6xl-runner
+    steps:
+      - name: Checkout DLC source
+        uses: actions/checkout@v5
+      
+      - name: ECR login
+        run: |
+          aws ecr get-login-password --region ${{ secrets.AWS_REGION }} | docker login --username AWS --password-stdin ${{ secrets.AWS_ACCOUNT_ID }}.dkr.ecr.${{ secrets.AWS_REGION }}.amazonaws.com
+ 
+      - name: Download image URI artifact
+        uses: actions/download-artifact@v4
+        with:
+          name: vllm-ec2-image-uri
+
+      - name: Resolve image URI for test
+        run: |
+          IMAGE_URI=$(cat image_uri.txt)
+          echo "Resolved image URI: $IMAGE_URI"
+          echo "IMAGE_URI=$IMAGE_URI" >> $GITHUB_ENV
+      
+      - name: Pull image
+        run: |
+          docker pull $IMAGE_URI
+      
+      - name: Checkout vLLM Tests
+        uses: actions/checkout@v5
+        with:
+          repository: vllm-project/vllm
+          ref: v0.11.0
+          path: vllm_source
+      
+      - name: Start container
+        run: |
+          CONTAINER_ID=$(docker run -d -it --rm --gpus=all --entrypoint /bin/bash \
+            -v ${HOME}/.cache/huggingface:/root/.cache/huggingface \
+            -v ${HOME}/.cache/vllm:/root/.cache/vllm \
+            -v ./vllm_source:/workdir --workdir /workdir \
+            -e HUGGING_FACE_HUB_TOKEN=${{ secrets.HUGGING_FACE_HUB_TOKEN }} \
+            ${IMAGE_URI})
+          echo "CONTAINER_ID=$CONTAINER_ID" >> $GITHUB_ENV
+      
+      - name: Setup for vLLM Test 
+        run: |
+          docker exec ${CONTAINER_ID} sh -c '
+            set -eux
+            uv pip install --system -r requirements/common.txt -r requirements/dev.txt --torch-backend=auto
+            uv pip install --system pytest pytest-asyncio
+            uv pip install --system -e tests/vllm_test_utils
+            uv pip install --system hf_transfer
+            mkdir src
+            mv vllm src/vllm
+          '
+      
+      - name: Run vLLM Tests
+        run: |
+          docker exec ${CONTAINER_ID} sh -c '
+          set -eux
+          nvidia-smi
+
+          # Platform Tests (CUDA) # 4min
+          cd /workdir/tests
+          pytest -v -s cuda/test_cuda_context.py
+          '
+      
+      - name: Cleanup container and images
+        if: always()
+        run: |
+          docker rm -f ${CONTAINER_ID} || true
+          docker image prune -a --force --filter "until=24h"
+          docker system df