[Feature] Support DeepEP normal & Redundant Experts on NPU (#9881)

.github/workflows/pr-test-npu.yml

@@ -127,12 +127,48 @@ jobs:
           cd test/srt
           python3 run_suite.py --suite per-commit-4-ascend-npu --timeout-per-file 3600
 
+  per-commit-16-ascend-a3:
+    if: (github.repository == 'sgl-project/sglang' || github.event_name == 'pull_request') &&
+        github.event.pull_request.draft == false
+    runs-on: linux-aarch64-a3-16
+    container:
+      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.2.rc1-a3-ubuntu22.04-py3.11
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Install dependencies
+        run: |
+          # speed up by using infra cache services
+          CACHING_URL="cache-service.nginx-pypi-cache.svc.cluster.local"
+          sed -Ei "s@(ports|archive).ubuntu.com@${CACHING_URL}:8081@g" /etc/apt/sources.list
+          pip config set global.index-url http://${CACHING_URL}/pypi/simple
+          pip config set global.trusted-host ${CACHING_URL}
+
+          bash scripts/ci/npu_ci_install_dependency.sh
+          # copy required file from our daily cache
+          cp ~/.cache/modelscope/hub/datasets/otavia/ShareGPT_Vicuna_unfiltered/ShareGPT_V3_unfiltered_cleaned_split.json /tmp
+          # copy download through proxy
+          curl -o /tmp/test.jsonl -L https://gh-proxy.test.osinfra.cn/https://raw.githubusercontent.com/openai/grade-school-math/master/grade_school_math/data/test.jsonl
+
+      - name: Run test
+        timeout-minutes: 90
+        env:
+          SGLANG_USE_MODELSCOPE: true
+          SGLANG_IS_IN_CI: true
+          HF_ENDPOINT: https://hf-mirror.com
+          TORCH_EXTENSIONS_DIR: /tmp/torch_extensions
+        run: |
+          cd test/srt
+          python3 run_suite.py --suite per-commit-16-ascend-a3 --timeout-per-file 5400
+
   pr-test-npu-finish:
     if: always()
     needs:
       - per-commit-1-ascend-npu
       - per-commit-2-ascend-npu
       - per-commit-4-ascend-npu
+      - per-commit-16-ascend-a3
     runs-on: ubuntu-latest
     steps:
       - name: Check all dependent job statuses

.github/workflows/release-docker-npu-nightly.yml

@@ -72,5 +72,6 @@ jobs:
           push: ${{ github.repository == 'sgl-project/sglang' && github.event_name != 'pull_request' }}
           provenance: false
           build-args: |
+            SGLANG_KERNEL_NPU_TAG=20250901
             CANN_VERSION=${{ matrix.cann_version }}
             DEVICE_TYPE=${{ matrix.device_type }}

.github/workflows/release-docker-npu.yml

@@ -54,8 +54,6 @@ jobs:
         run: |
           version=$(cat python/sglang/version.py | cut -d'"' -f2)
           echo "TAG=lmsysorg/sglang:v$version-cann${{ matrix.cann_version }}-${{ matrix.device_type }}" >> $GITHUB_OUTPUT
-          kernel_tag=$(curl -s https://api.github.com/repos/sgl-project/sgl-kernel-npu/tags | jq -r '.[0].name')
-          echo "KERNEL_NPU_TAG=${kernel_tag}" >> $GITHUB_OUTPUT
 
       - name: Build and push Docker image
         id: build-and-push
@@ -70,6 +68,6 @@ jobs:
           push: ${{ github.repository == 'sgl-project/sglang' && github.event_name != 'pull_request' }}
           provenance: false
           build-args: |
-            SGLANG_KERNEL_NPU_TAG=${{ steps.get_version.outputs.KERNEL_NPU_TAG }}
+            SGLANG_KERNEL_NPU_TAG=20250901
             CANN_VERSION=${{ matrix.cann_version }}
             DEVICE_TYPE=${{ matrix.device_type }}