Fix sgl-deep-gemm release workflow (#24385)

Co-authored-by: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

.github/workflows/release-whl-deepgemm.yml

@@ -33,7 +33,7 @@ jobs:
       (github.event.inputs.target == 'all' || github.event.inputs.target == 'cu129')
     strategy:
       matrix:
-        python-version: ["3.10"]
+        python-version: ["3.12"]
         cuda-version: ["12.9"]
         arch: [x86_64, aarch64]
         include:
@@ -65,14 +65,9 @@ jobs:
 
       - name: Build wheel
         run: |
-          chmod +x ./scripts/build_sgl_deep_gemm.sh
+          chmod +x ./scripts/build_sgl_deep_gemm.sh ./scripts/rename_sgl_deep_gemm_whl.sh
           ./scripts/build_sgl_deep_gemm.sh "${{ matrix.python-version }}" "${{ matrix.cuda-version }}" "${{ github.workspace }}/DeepGEMM" "${{ matrix.arch }}"
 
-      - name: Rename wheel
-        run: |
-          chmod +x ./scripts/rename_sgl_deep_gemm_whl.sh
-          ./scripts/rename_sgl_deep_gemm_whl.sh DeepGEMM/dist cu129 ${{ matrix.arch }}
-
       - name: Upload artifacts
         uses: actions/upload-artifact@v4
         with:
@@ -124,7 +119,7 @@ jobs:
       (github.event.inputs.target == 'all' || github.event.inputs.target == 'cu130')
     strategy:
       matrix:
-        python-version: ["3.10"]
+        python-version: ["3.12"]
         cuda-version: ["13.0"]
         arch: [x86_64, aarch64]
         include:
@@ -149,6 +144,11 @@ jobs:
           path: DeepGEMM
           submodules: recursive
 
+      - name: Set up Python ${{ matrix.python-version }}
+        uses: actions/setup-python@v5
+        with:
+          python-version: ${{ matrix.python-version }}
+
       - name: Set wheel version
         run: |
           echo -n "${{ inputs.version }}" > DeepGEMM/sgl_deep_gemm/VERSION
@@ -156,41 +156,56 @@ jobs:
 
       - name: Build wheel
         run: |
-          chmod +x ./scripts/build_sgl_deep_gemm.sh
+          chmod +x ./scripts/build_sgl_deep_gemm.sh ./scripts/rename_sgl_deep_gemm_whl.sh
           ./scripts/build_sgl_deep_gemm.sh "${{ matrix.python-version }}" "${{ matrix.cuda-version }}" "${{ github.workspace }}/DeepGEMM" "${{ matrix.arch }}"
 
-      - name: Rename wheel
-        run: |
-          chmod +x ./scripts/rename_sgl_deep_gemm_whl.sh
-          ./scripts/rename_sgl_deep_gemm_whl.sh DeepGEMM/dist cu130 ${{ matrix.arch }}
-
-      - name: Strip +cu130 local version for PyPI upload
-        working-directory: DeepGEMM
-        run: |
-          set -eux
-          pip install wheel
-          mkdir -p dist-pypi
-          for w in dist/*.whl; do
-            tmp=$(mktemp -d)
-            python3 -m wheel unpack "$w" --dest "$tmp"
-            unpacked=$(find "$tmp" -mindepth 1 -maxdepth 1 -type d | head -1)
-            info=$(find "$unpacked" -maxdepth 1 -type d -name "*.dist-info" | head -1)
-            meta="$info/METADATA"
-            orig=$(grep '^Version:' "$meta" | head -1 | sed 's/^Version:[[:space:]]*//')
-            new=$(echo "$orig" | sed 's/+cu[0-9]\+$//')
-            if [ "$orig" != "$new" ]; then
-              sed -i "s/^Version:.*/Version: ${new}/" "$meta"
-              old_base=$(basename "$info")
-              new_base="${old_base/${orig}/${new}}"
-              mv "$info" "$(dirname "$info")/${new_base}"
-            fi
-            python3 -m wheel pack "$unpacked" --dest-dir dist-pypi
-            rm -rf "$tmp"
-          done
-          ls -lh dist-pypi/
-
       - name: Upload to PyPI
         working-directory: DeepGEMM
         run: |
           pip install twine
           python3 -m twine upload --skip-existing dist-pypi/* -u __token__ -p ${{ secrets.SGL_DEEP_GEMM_PYPI_TOKEN }}
+
+      - name: Upload artifacts
+        uses: actions/upload-artifact@v4
+        with:
+          name: deepgemm-wheel-cuda${{ matrix.cuda-version }}-${{ matrix.arch }}
+          path: DeepGEMM/dist/*.whl
+
+  release-cu130:
+    needs: build-cu130-matrix
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - name: Download artifacts
+        uses: actions/download-artifact@v4
+        with:
+          path: dist/
+          merge-multiple: true
+          pattern: deepgemm-wheel-cuda13.0-*
+
+      - name: Release
+        uses: softprops/action-gh-release@v2
+        with:
+          tag_name: v${{ inputs.version }}
+          repository: sgl-project/whl
+          token: ${{ secrets.GH_PAT_FOR_WHL_RELEASE }}
+          files: |
+            dist/*
+
+      - name: Clone wheel index
+        run: git clone https://oauth2:${WHL_TOKEN}@github.com/sgl-project/whl.git sgl-whl
+        env:
+          WHL_TOKEN: ${{ secrets.GH_PAT_FOR_WHL_RELEASE }}
+
+      - name: Update wheel index
+        run: python3 scripts/update_deepgemm_whl_index.py --cuda 130
+
+      - name: Push wheel index
+        run: |
+          cd sgl-whl
+          git config --local user.name "sglang-bot"
+          git config --local user.email "sglangbot@gmail.com"
+          git add -A
+          git commit -m "update sgl-deep-gemm whl index for v${{ inputs.version }}"
+          git push

docker/sgl-deep-gemm.Dockerfile

@@ -5,8 +5,8 @@ FROM ${BASE_IMG}:cuda${CUDA_VERSION}
 
 ARG ARCH=x86_64
 ARG CUDA_VERSION=13.0
-ARG PYTHON_VERSION=3.10
+ARG PYTHON_VERSION=3.12
-ARG PYTHON_TAG=cp310-cp310
+ARG PYTHON_TAG=cp312-cp312
 ARG TORCH_VER=2.11.0
 ARG TVM_FFI_VER=0.1.9
 ARG PIP_DEFAULT_INDEX=https://pypi.python.org/simple

scripts/build_sgl_deep_gemm.sh

@@ -7,7 +7,10 @@
 #   DEEPGEMM_SRC:   path to a checkout of sgl-project/DeepGEMM
 #   ARCH:           x86_64 (default) or aarch64
 #
-# The wheel is written into <DEEPGEMM_SRC>/dist by build_sgl_deep_gemm.sh.
+# Writes:
+#   <DEEPGEMM_SRC>/dist/      — wheel(s) tagged +cu129 / +cu130 and manylinux
+#   <DEEPGEMM_SRC>/dist-pypi/ — cu130 only: same wheel(s) with +cu130 stripped
+#                              (PyPI rejects local-version segments)
 set -ex
 
 if [ $# -lt 3 ]; then
@@ -20,6 +23,15 @@ CUDA_VERSION="$2"
 DEEPGEMM_SRC="$(cd "$3" && pwd)"
 ARCH="${4:-$(uname -i)}"
 
+case "${CUDA_VERSION}" in
+  13.0) CU_TAG=cu130 ;;
+  12.9) CU_TAG=cu129 ;;
+  *)
+    echo "Unsupported CUDA_VERSION: ${CUDA_VERSION}" >&2
+    exit 1
+    ;;
+esac
+
 if [ "${ARCH}" = "aarch64" ]; then
   BASE_IMG="pytorch/manylinuxaarch64-builder"
 else
@@ -31,12 +43,14 @@ PY_TAG="cp${PYTHON_VERSION//.}-cp${PYTHON_VERSION//.}"
 SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
 REPO_ROOT="$(cd "${SCRIPT_DIR}/.." && pwd)"
 DOCKERFILE="${REPO_ROOT}/docker/sgl-deep-gemm.Dockerfile"
+RENAME_SCRIPT="${SCRIPT_DIR}/rename_sgl_deep_gemm_whl.sh"
 
 DEPS_TAG="sgl-deep-gemm-deps:cuda${CUDA_VERSION}-${PY_TAG}-${ARCH}"
 
 echo "----------------------------------------"
 echo "PYTHON_VERSION: ${PYTHON_VERSION}"
 echo "CUDA_VERSION:   ${CUDA_VERSION}"
+echo "CU_TAG:         ${CU_TAG}"
 echo "ARCH:           ${ARCH}"
 echo "BASE_IMG:       ${BASE_IMG}"
 echo "DEEPGEMM_SRC:   ${DEEPGEMM_SRC}"
@@ -55,6 +69,7 @@ docker build \
 
 mkdir -p "${DEEPGEMM_SRC}/dist"
 
+# 1) Build the wheel inside the deps container.
 docker run --rm \
   --network=host \
   -v "${DEEPGEMM_SRC}:/deepgemm" \
@@ -62,5 +77,49 @@ docker run --rm \
   "${DEPS_TAG}" \
   bash build_sgl_deep_gemm.sh
 
-echo "Wheels written to ${DEEPGEMM_SRC}/dist:"
+# 2) Rename inside the same image so we have a working pip / wheel CLI and can
+#    rewrite the root-owned wheel files written by the build container above.
+docker run --rm \
+  -v "${DEEPGEMM_SRC}:/deepgemm" \
+  -v "${RENAME_SCRIPT}:/rename_sgl_deep_gemm_whl.sh:ro" \
+  -w /deepgemm \
+  "${DEPS_TAG}" \
+  bash /rename_sgl_deep_gemm_whl.sh dist "${CU_TAG}" "${ARCH}"
+
+# 3) cu130 only: produce a sibling dist-pypi/ with the +cu130 local-version
+#    stripped (PyPI rejects local versions).
+if [ "${CU_TAG}" = "cu130" ]; then
+  docker run --rm \
+    -v "${DEEPGEMM_SRC}:/deepgemm" \
+    -w /deepgemm \
+    "${DEPS_TAG}" \
+    bash -c '
+set -eux
+mkdir -p dist-pypi
+for w in dist/*.whl; do
+  tmp=$(mktemp -d)
+  python3 -m wheel unpack "$w" --dest "$tmp"
+  unpacked=$(find "$tmp" -mindepth 1 -maxdepth 1 -type d | head -1)
+  info=$(find "$unpacked" -maxdepth 1 -type d -name "*.dist-info" | head -1)
+  meta="$info/METADATA"
+  orig=$(grep "^Version:" "$meta" | head -1 | sed "s/^Version:[[:space:]]*//")
+  new=$(echo "$orig" | sed "s/+cu[0-9]\+$//")
+  if [ "$orig" != "$new" ]; then
+    sed -i "s/^Version:.*/Version: ${new}/" "$meta"
+    old_base=$(basename "$info")
+    new_base="${old_base/${orig}/${new}}"
+    mv "$info" "$(dirname "$info")/${new_base}"
+  fi
+  python3 -m wheel pack "$unpacked" --dest-dir dist-pypi
+  rm -rf "$tmp"
+done
+ls -lh dist-pypi/
+'
+fi
+
+echo "Wheels in ${DEEPGEMM_SRC}/dist:"
 ls -lh "${DEEPGEMM_SRC}/dist"/*.whl 2>/dev/null || true
+if [ "${CU_TAG}" = "cu130" ]; then
+  echo "PyPI-ready wheels in ${DEEPGEMM_SRC}/dist-pypi:"
+  ls -lh "${DEEPGEMM_SRC}/dist-pypi"/*.whl 2>/dev/null || true
+fi