Update to flashinfer-python==0.2.12 and disable AOT compile for non-release image (#23129)

Signed-off-by: mgoin <mgoin64@gmail.com>

.buildkite/release-pipeline.yaml

@@ -68,7 +68,7 @@ steps:
       queue: cpu_queue_postmerge
     commands:
       - "aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin public.ecr.aws/q9t5s3a7"
-      - "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=12.8.1 --build-arg INSTALL_KV_CONNECTORS=true --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT --target vllm-openai --progress plain -f docker/Dockerfile ."
+      - "DOCKER_BUILDKIT=1 docker build --build-arg max_jobs=16 --build-arg USE_SCCACHE=1 --build-arg GIT_REPO_CHECK=1 --build-arg CUDA_VERSION=12.8.1 --build-arg FLASHINFER_AOT_COMPILE=true --build-arg INSTALL_KV_CONNECTORS=true --tag public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT --target vllm-openai --progress plain -f docker/Dockerfile ."
       - "docker push public.ecr.aws/q9t5s3a7/vllm-release-repo:$BUILDKITE_COMMIT"
 
   - label: "Annotate release workflow"

docker/Dockerfile

@@ -372,31 +372,45 @@ RUN --mount=type=bind,from=build,src=/workspace/dist,target=/vllm-workspace/dist
 
 # Install FlashInfer from source
 ARG FLASHINFER_GIT_REPO="https://github.com/flashinfer-ai/flashinfer.git"
-# Keep this in sync with https://github.com/vllm-project/vllm/blob/main/requirements/cuda.txt
+# Keep this in sync with "flashinfer" extra in setup.py
-# We use `--force-reinstall --no-deps` to avoid issues with the existing FlashInfer wheel.
+ARG FLASHINFER_GIT_REF="v0.2.12"
-ARG FLASHINFER_GIT_REF="v0.2.11"
+# Flag to control whether to compile FlashInfer AOT kernels
+# Set to "true" to enable AOT compilation:
+# docker build --build-arg FLASHINFER_AOT_COMPILE=true ...
+ARG FLASHINFER_AOT_COMPILE=false
 RUN --mount=type=cache,target=/root/.cache/uv bash - <<'BASH'
   . /etc/environment
     git clone --depth 1 --recursive --shallow-submodules \
         --branch ${FLASHINFER_GIT_REF} \
         ${FLASHINFER_GIT_REPO} flashinfer
-    # Exclude CUDA arches for older versions (11.x and 12.0-12.7)
-    # TODO: Update this to allow setting TORCH_CUDA_ARCH_LIST as a build arg.
-    if [[ "${CUDA_VERSION}" == 11.* ]]; then
-        FI_TORCH_CUDA_ARCH_LIST="7.5 8.0 8.9"
-    elif [[ "${CUDA_VERSION}" == 12.[0-7]* ]]; then
-        FI_TORCH_CUDA_ARCH_LIST="7.5 8.0 8.9 9.0a"
-    else
-        # CUDA 12.8+ supports 10.0a and 12.0
-        FI_TORCH_CUDA_ARCH_LIST="7.5 8.0 8.9 9.0a 10.0a 12.0"
-    fi
-    echo "🏗️  Building FlashInfer for arches: ${FI_TORCH_CUDA_ARCH_LIST}"
-    # Needed to build AOT kernels
     pushd flashinfer
-        TORCH_CUDA_ARCH_LIST="${FI_TORCH_CUDA_ARCH_LIST}" \
+        if [ "${FLASHINFER_AOT_COMPILE}" = "true" ]; then
-            python3 -m flashinfer.aot
+            # Exclude CUDA arches for older versions (11.x and 12.0-12.7)
-        TORCH_CUDA_ARCH_LIST="${FI_TORCH_CUDA_ARCH_LIST}" \
+            # TODO: Update this to allow setting TORCH_CUDA_ARCH_LIST as a build arg.
-            uv pip install --system --no-build-isolation --force-reinstall --no-deps .
+            if [[ "${CUDA_VERSION}" == 11.* ]]; then
+                FI_TORCH_CUDA_ARCH_LIST="7.5 8.0 8.9"
+            elif [[ "${CUDA_VERSION}" == 12.[0-7]* ]]; then
+                FI_TORCH_CUDA_ARCH_LIST="7.5 8.0 8.9 9.0a"
+            else
+                # CUDA 12.8+ supports 10.0a and 12.0
+                FI_TORCH_CUDA_ARCH_LIST="7.5 8.0 8.9 9.0a 10.0a 12.0"
+            fi
+            echo "🏗️  Installing FlashInfer with AOT compilation for arches: ${FI_TORCH_CUDA_ARCH_LIST}"
+            # Build AOT kernels
+            TORCH_CUDA_ARCH_LIST="${FI_TORCH_CUDA_ARCH_LIST}" \
+                python3 -m flashinfer.aot
+            # Install with no-build-isolation since we already built AOT kernels
+            TORCH_CUDA_ARCH_LIST="${FI_TORCH_CUDA_ARCH_LIST}" \
+                uv pip install --system --no-build-isolation . \
+                --extra-index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.')
+            # Download pre-compiled cubins
+            TORCH_CUDA_ARCH_LIST="${FI_TORCH_CUDA_ARCH_LIST}" \
+                python3 -m flashinfer --download-cubin || echo "WARNING: Failed to download flashinfer cubins."
+        else
+            echo "🏗️  Installing FlashInfer without AOT compilation in JIT mode"
+            uv pip install --system . \
+                --extra-index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.')
+        fi
     popd
     rm -rf flashinfer
 BASH

setup.py

@@ -685,7 +685,7 @@ setup(
                   "mistral_common[audio]"],  # Required for audio processing
         "video": [],  # Kept for backwards compatibility
         # FlashInfer should be updated together with the Dockerfile
-        "flashinfer": ["flashinfer-python==0.2.11"],
+        "flashinfer": ["flashinfer-python==0.2.12"],
     },
     cmdclass=cmdclass,
     package_data=package_data,