Merge branch 'main' into main

Author: xieminghe1

.github/CODEOWNERS

@@ -25,6 +25,7 @@
 /python/sglang/srt/mem_cache @merrymercy @Ying1123 @hnyls2002 @xiezhq-hermann @hanming-lu @yizhang2077
 /python/sglang/srt/model_executor @merrymercy @Ying1123 @hnyls2002 @Fridge003 @ispobock
 /python/sglang/srt/model_executor/piecewise_cuda_graph_runner.py @hebiao064
+/python/sglang/srt/models/deepseek_v2.py @fzyzcjy @zhyncs @ispobock @ch-wan @merrymercy @Fridge003
 /python/sglang/srt/multimodal @mickqian @JustinTong0323 @yhyang201 @yuan-luo
 /python/sglang/srt/speculative @Ying1123 @merrymercy @hnyls2002
 /sgl-kernel @zhyncs @ispobock @BBuf @yizhang2077 @merrymercy @FlamingoPg @HaiShaw

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

@@ -23,7 +23,7 @@ jobs:
       matrix:
         part: [0, 1]
     container:
-      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc1-a3-ubuntu22.04-py3.11
+      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc2-a3-ubuntu22.04-py3.11
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
@@ -69,7 +69,7 @@ jobs:
       matrix:
         part: [0]
     container:
-      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc1-a3-ubuntu22.04-py3.11
+      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc2-a3-ubuntu22.04-py3.11
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
@@ -115,7 +115,7 @@ jobs:
       matrix:
         part: [0]
     container:
-      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc1-a3-ubuntu22.04-py3.11
+      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc2-a3-ubuntu22.04-py3.11
     steps:
       - name: Checkout code
         uses: actions/checkout@v4

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

@@ -45,7 +45,7 @@ jobs:
     if: needs.check-changes.outputs.main_package == 'true'
     runs-on: linux-arm64-npu-1
     container:
-      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc1-910b-ubuntu22.04-py3.11
+      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc2-910b-ubuntu22.04-py3.11
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
@@ -88,7 +88,7 @@ jobs:
       matrix:
         part: [0, 1, 2]
     container:
-      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc1-910b-ubuntu22.04-py3.11
+      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc2-910b-ubuntu22.04-py3.11
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
@@ -127,7 +127,7 @@ jobs:
     if: needs.check-changes.outputs.main_package == 'true'
     runs-on: linux-arm64-npu-4
     container:
-      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc1-910b-ubuntu22.04-py3.11
+      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc2-910b-ubuntu22.04-py3.11
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
@@ -170,7 +170,7 @@ jobs:
       matrix:
         part: [0, 1]
     container:
-      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc1-a3-ubuntu22.04-py3.11
+      image: swr.cn-southwest-2.myhuaweicloud.com/base_image/ascend-ci/cann:8.3.rc2-a3-ubuntu22.04-py3.11
     steps:
       - name: Checkout code
         uses: actions/checkout@v4

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

@@ -330,8 +330,31 @@ jobs:
           docker rm oracle-db || true
 
 
+  docker-build-test:
+    if: |
+      github.event_name != 'pull_request' ||
+      (github.event.action != 'labeled' && contains(github.event.pull_request.labels.*.name, 'run-ci')) ||
+      (github.event.action == 'labeled' && github.event.label.name == 'run-ci')
+    runs-on: ubuntu-24.04
+    steps:
+      - name: Checkout repository
+        uses: actions/checkout@v4
+
+      - name: Set up Docker Buildx
+        uses: docker/setup-buildx-action@v3
+
+      - name: Build Docker image (no push)
+        uses: docker/build-push-action@v5
+        with:
+          context: .
+          file: docker/gateway.Dockerfile
+          push: false
+          tags: sglang-router:test
+          cache-from: type=gha
+          cache-to: type=gha,mode=max
+
   finish:
-    needs: [maturin-build-test, router-unit-tests, router-http-tests, router-grpc-response-api-tests]
+    needs: [maturin-build-test, router-unit-tests, router-http-tests, router-grpc-response-api-tests, docker-build-test]
     runs-on: ubuntu-latest
     steps:
       - name: Finish

.github/workflows/pr-test.yml

@@ -36,6 +36,7 @@ jobs:
       main_package: ${{ steps.filter.outputs.main_package || steps.scheduled.outputs.main_package }}
       sgl_kernel: ${{ steps.filter.outputs.sgl_kernel || steps.scheduled.outputs.sgl_kernel }}
       multimodal_gen: ${{ steps.filter.outputs.multimodal_gen || steps.scheduled.outputs.multimodal_gen }}
+      max_parallel: ${{ steps.set-parallel.outputs.max_parallel }}
     steps:
       - name: Checkout code
         uses: actions/checkout@v4
@@ -68,6 +69,17 @@ jobs:
           echo "sgl_kernel=false" >> $GITHUB_OUTPUT
           echo "multimodal_gen=true" >> $GITHUB_OUTPUT
 
+      - name: Set max-parallel based on high-priority label
+        id: set-parallel
+        run: |
+          if [[ "${{ github.event_name }}" == "pull_request" && "${{ contains(github.event.pull_request.labels.*.name, 'high priority') }}" == "true" ]]; then
+            echo "max_parallel=15" >> $GITHUB_OUTPUT
+            echo "High priority PR detected, setting max_parallel to 15"
+          else
+            echo "max_parallel=8" >> $GITHUB_OUTPUT
+            echo "Using default max_parallel of 8"
+          fi
+
       - name: Show filter results in summary (table)
         run: |
           {
@@ -78,6 +90,7 @@ jobs:
             echo "| main_package   | ${{ steps.filter.outputs.main_package || steps.scheduled.outputs.main_package }} |"
             echo "| sgl_kernel     | ${{ steps.filter.outputs.sgl_kernel || steps.scheduled.outputs.sgl_kernel }} |"
             echo "| multimodal_gen | ${{ steps.filter.outputs.multimodal_gen || steps.scheduled.outputs.multimodal_gen }} |"
+            echo "| max_parallel   | ${{ steps.set-parallel.outputs.max_parallel }} |"
           } >> $GITHUB_STEP_SUMMARY
 
   # =============================================== PR Gate ====================================================
@@ -382,6 +395,46 @@ jobs:
           # temporarily put backend-independent cpu tests here
           python3 run_suite.py --hw cpu --suite default
 
+  stage-b-test-small-1-gpu:
+    needs: [check-changes, call-gate, stage-a-test-1, sgl-kernel-build-wheels]
+    if: |
+      always() &&
+      (
+        (inputs.target_stage == 'stage-b-test-small-1-gpu') ||
+        (
+          !inputs.target_stage &&
+          (github.event_name == 'schedule' || (!failure() && !cancelled())) &&
+          ((needs.check-changes.outputs.main_package == 'true') || (needs.check-changes.outputs.sgl_kernel == 'true'))
+        )
+      )
+    runs-on: 1-gpu-runner
+    env:
+      RUNNER_LABELS: 1-gpu-runner
+    strategy:
+      fail-fast: false
+      matrix:
+        partition: [0, 1, 2, 3]
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+
+      - name: Download artifacts
+        if: needs.check-changes.outputs.sgl_kernel == 'true'
+        uses: actions/download-artifact@v4
+        with:
+          path: sgl-kernel/dist/
+          merge-multiple: true
+          pattern: wheel-python3.10-cuda12.9
+
+      - name: Install dependencies
+        run: |
+          CUSTOM_BUILD_SGL_KERNEL=${{needs.check-changes.outputs.sgl_kernel}} bash scripts/ci/ci_install_dependency.sh
+
+      - name: Run test
+        timeout-minutes: 30
+        run: |
+          cd test/
+          python3 run_suite.py --hw cuda --suite stage-b-test-small-1-gpu --auto-partition-id ${{ matrix.partition }} --auto-partition-size 4
 
   multimodal-gen-test-1-gpu:
     needs: [check-changes, call-gate, sgl-kernel-build-wheels]
@@ -529,7 +582,7 @@ jobs:
       RUNNER_LABELS: 1-gpu-runner
     strategy:
       fail-fast: false
-      max-parallel: 8
+      max-parallel: ${{ fromJson(needs.check-changes.outputs.max_parallel) }}
       matrix:
         part: [0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14]
     steps:
@@ -1143,7 +1196,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        part: [0, 1]
+        part: [0, 1, 2]
 
     steps:
       - name: Checkout code
@@ -1165,7 +1218,7 @@ jobs:
         timeout-minutes: 30
         run: |
           cd test/srt
-          python3 run_suite.py --suite per-commit-4-gpu-b200 --auto-partition-id ${{ matrix.part }}  --auto-partition-size 2 --timeout-per-file 1800
+          python3 run_suite.py --suite per-commit-4-gpu-b200 --auto-partition-id ${{ matrix.part }}  --auto-partition-size 3 --timeout-per-file 1800
 
   unit-test-backend-4-gpu-gb200:
     needs: [check-changes, call-gate, unit-test-backend-2-gpu, sgl-kernel-build-wheels-arm]
@@ -1209,11 +1262,14 @@ jobs:
   unit-test-backend-8-gpu-b200:
     needs: [check-changes, call-gate, unit-test-backend-2-gpu]
     if: |
-      (inputs.target_stage == 'unit-test-backend-8-gpu-b200') ||
+      always() &&
       (
-        always() &&
-        (github.event_name == 'schedule' || (!failure() && !cancelled())) &&
-        ((needs.check-changes.outputs.main_package == 'true') || (needs.check-changes.outputs.sgl_kernel == 'true'))
+        (inputs.target_stage == 'unit-test-backend-8-gpu-b200') ||
+        (
+          !inputs.target_stage &&
+          (github.event_name == 'schedule' || (!failure() && !cancelled())) &&
+          ((needs.check-changes.outputs.main_package == 'true') || (needs.check-changes.outputs.sgl_kernel == 'true'))
+        )
       )
     runs-on: 8-gpu-b200
     env:
@@ -1226,11 +1282,11 @@ jobs:
         run: |
           IS_BLACKWELL=1 bash scripts/ci/ci_install_dependency.sh
 
-      - name: Run test
-        timeout-minutes: 45
-        run: |
-          cd test/srt
-          python3 run_suite.py --suite per-commit-8-gpu-b200 --timeout-per-file 1800
+      # - name: Run test
+      #   timeout-minutes: 45
+      #   run: |
+      #     cd test/srt
+      #     python3 run_suite.py --suite per-commit-8-gpu-b200 --timeout-per-file 1800
 
   pr-test-finish:
     needs:
@@ -1248,6 +1304,7 @@ jobs:
         multimodal-gen-test-2-gpu,
 
         stage-a-test-1,
+        stage-b-test-small-1-gpu,
         quantization-test,
         unit-test-backend-1-gpu,
         unit-test-backend-2-gpu,
@@ -1264,7 +1321,7 @@ jobs:
         unit-test-deepep-8-gpu,
         unit-test-backend-4-gpu-b200,
         unit-test-backend-4-gpu-gb200,
-        unit-test-backend-8-gpu-b200,
+        # unit-test-backend-8-gpu-b200,  # Moved to nightly - large models only
       ]
     if: always()
     runs-on: ubuntu-latest

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

@@ -4,7 +4,7 @@ on:
     branches:
       - main
     paths:
-      - "sgl-model-gateway/bindings/python/sglang_router/version.py"
+      - sgl-model-gateway/bindings/python/pyproject.toml
   workflow_dispatch:
 
 jobs:

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

@@ -19,7 +19,7 @@ jobs:
     runs-on: ubuntu-22.04-arm
     strategy:
       matrix:
-        cann_version: ["8.3.rc1"]
+        cann_version: ["8.3.rc2"]
         device_type: ["910b", "a3"]
     steps:
       - name: Checkout repository
@@ -73,6 +73,6 @@ jobs:
           push: ${{ github.repository == 'sgl-project/sglang' && github.event_name != 'pull_request' }}
           provenance: false
           build-args: |
-            SGLANG_KERNEL_NPU_TAG=20251128
+            SGLANG_KERNEL_NPU_TAG=20251206
             CANN_VERSION=${{ matrix.cann_version }}
             DEVICE_TYPE=${{ matrix.device_type }}

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

@@ -17,7 +17,7 @@ jobs:
     runs-on: ubuntu-22.04-arm
     strategy:
       matrix:
-        cann_version: ["8.3.rc1"]
+        cann_version: ["8.3.rc2"]
         device_type: ["910b", "a3"]
     steps:
       - name: Checkout repository
@@ -70,6 +70,6 @@ jobs:
           push: ${{ github.repository == 'sgl-project/sglang' && github.event_name != 'pull_request' }}
           provenance: false
           build-args: |
-            SGLANG_KERNEL_NPU_TAG=20251128
+            SGLANG_KERNEL_NPU_TAG=20251206
             CANN_VERSION=${{ matrix.cann_version }}
             DEVICE_TYPE=${{ matrix.device_type }}

Makefile

@@ -33,7 +33,7 @@ Add [performance optimization options](#performance-optimization-options) as nee
 
 ```bash
 # Installation
-pip install "sglang[all]>=0.5.6"
+pip install "sglang[all]>=0.5.6.post1"
 
 # Launch
 python3 -m sglang.launch_server --model deepseek-ai/DeepSeek-V3 --tp 8 --trust-remote-code