Merge branch 'main' into mistral-tool-parser-streaming-update

Author: avigny

.buildkite/scripts/scheduled_integration_test/deepseek_v2_lite_ep_async_eplb.sh

@@ -0,0 +1,73 @@
+#!/usr/bin/env bash
+set -euxo pipefail
+
+# args: [THRESHOLD] [NUM_QUESTIONS] [START_PORT]
+THRESHOLD=${1:-0.25}
+NUM_Q=${2:-1319}
+PORT=${3:-8030}
+OUT_DIR=${OUT_DIR:-/tmp/vllm-scheduled}
+mkdir -p "${OUT_DIR}"
+
+wait_for_server() {
+  local port=$1
+  timeout 600 bash -c '
+    until curl -sf "http://127.0.0.1:'"$port"'/health" > /dev/null; do
+      sleep 1
+    done'
+}
+
+MODEL="deepseek-ai/DeepSeek-V2-lite"
+
+# Set BACKENDS based on platform
+if command -v rocm-smi &> /dev/null || [[ -d /opt/rocm ]] || [[ -n "${ROCM_PATH:-}" ]]; then
+  # ROCm platform
+  BACKENDS=("allgather_reducescatter")
+  # Disable MOE padding for ROCm since it is causing eplb to fail
+  export VLLM_ROCM_MOE_PADDING=0
+else
+  # Non-ROCm platform (CUDA/other)
+  BACKENDS=("deepep_high_throughput" "deepep_low_latency")
+fi
+
+cleanup() {
+  if [[ -n "${SERVER_PID:-}" ]] && kill -0 "${SERVER_PID}" 2>/dev/null; then
+    kill "${SERVER_PID}" 2>/dev/null || true
+    for _ in {1..20}; do
+      kill -0 "${SERVER_PID}" 2>/dev/null || break
+      sleep 0.5
+    done
+    kill -9 "${SERVER_PID}" 2>/dev/null || true
+  fi
+}
+trap cleanup EXIT
+
+for BACK in "${BACKENDS[@]}"; do
+  VLLM_DEEP_GEMM_WARMUP=skip \
+  VLLM_ALL2ALL_BACKEND=$BACK \
+  vllm serve "$MODEL" \
+    --enforce-eager \
+    --tensor-parallel-size 2 \
+    --data-parallel-size 2 \
+    --enable-expert-parallel \
+    --enable-eplb \
+    --eplb-config '{"window_size":200,"step_interval":600,"use_async":true}' \
+    --trust-remote-code \
+    --max-model-len 2048 \
+    --port $PORT &
+  SERVER_PID=$!
+  wait_for_server $PORT
+
+  TAG=$(echo "$MODEL" | tr '/: \\n' '_____')
+  OUT="${OUT_DIR}/${TAG}_${BACK}_async_eplb.json"
+  python3 tests/evals/gsm8k/gsm8k_eval.py --host http://127.0.0.1 --port $PORT --num-questions ${NUM_Q} --save-results ${OUT}
+  python3 - <<PY
+import json; acc=json.load(open('${OUT}'))['accuracy']
+print(f"${MODEL} ${BACK}: accuracy {acc:.3f}")
+assert acc >= ${THRESHOLD}, f"${MODEL} ${BACK} accuracy {acc}"
+PY
+
+  cleanup
+  SERVER_PID=
+  sleep 1
+  PORT=$((PORT+1))
+done

.buildkite/scripts/scheduled_integration_test/deepseek_v2_lite_ep_eplb.sh

@@ -50,6 +50,7 @@ for BACK in "${BACKENDS[@]}"; do
     --data-parallel-size 2 \
     --enable-expert-parallel \
     --enable-eplb \
+    --eplb-config '{"window_size":200,"step_interval":600}' \
     --trust-remote-code \
     --max-model-len 2048 \
     --port $PORT &

.buildkite/scripts/scheduled_integration_test/qwen3_next_mtp_async_eplb.sh

@@ -0,0 +1,74 @@
+#!/usr/bin/env bash
+set -euxo pipefail
+
+# args: [THRESHOLD] [NUM_QUESTIONS] [START_PORT]
+THRESHOLD=${1:-0.25}
+NUM_Q=${2:-1319}
+PORT=${3:-8040}
+OUT_DIR=${OUT_DIR:-/tmp/vllm-scheduled}
+mkdir -p "${OUT_DIR}"
+
+wait_for_server() {
+  local port=$1
+  timeout 600 bash -c '
+    until curl -sf "http://127.0.0.1:'"$port"'/health" > /dev/null; do
+      sleep 1
+    done'
+}
+
+MODEL="Qwen/Qwen3-Next-80B-A3B-Instruct"
+
+# Set BACKENDS based on platform
+if command -v rocm-smi &> /dev/null || [[ -d /opt/rocm ]] || [[ -n "${ROCM_PATH:-}" ]]; then
+  # ROCm platform
+  BACKENDS=("allgather_reducescatter")
+  # Disable MOE padding for ROCm since it is causing eplb to fail
+  export VLLM_ROCM_MOE_PADDING=0
+else
+  # Non-ROCm platform (CUDA/other)
+  BACKENDS=("deepep_high_throughput" "deepep_low_latency")
+fi
+
+cleanup() {
+  if [[ -n "${SERVER_PID:-}" ]] && kill -0 "${SERVER_PID}" 2>/dev/null; then
+    kill "${SERVER_PID}" 2>/dev/null || true
+    for _ in {1..20}; do
+      kill -0 "${SERVER_PID}" 2>/dev/null || break
+      sleep 0.5
+    done
+    kill -9 "${SERVER_PID}" 2>/dev/null || true
+  fi
+}
+trap cleanup EXIT
+
+for BACK in "${BACKENDS[@]}"; do
+  VLLM_DEEP_GEMM_WARMUP=skip \
+  VLLM_ALL2ALL_BACKEND=$BACK \
+  vllm serve "$MODEL" \
+    --enforce-eager \
+    --tensor-parallel-size 4 \
+    --enable-expert-parallel \
+    --enable-eplb \
+    --eplb-config '{"window_size":200,"step_interval":600,"use_async":true}' \
+    --speculative-config '{"method":"qwen3_next_mtp","num_speculative_tokens":1}' \
+    --trust-remote-code \
+    --max-model-len 2048 \
+    --gpu-memory-utilization 0.9 \
+    --port $PORT &
+  SERVER_PID=$!
+  wait_for_server $PORT
+
+  TAG=$(echo "$MODEL" | tr '/: \\n' '_____')
+  OUT="${OUT_DIR}/${TAG}_${BACK}.json"
+  python3 tests/evals/gsm8k/gsm8k_eval.py --host http://127.0.0.1 --port $PORT --num-questions ${NUM_Q} --save-results ${OUT}
+  python3 - <<PY
+import json; acc=json.load(open('${OUT}'))['accuracy']
+print(f"${MODEL} ${BACK}: accuracy {acc:.3f}")
+assert acc >= ${THRESHOLD}, f"${MODEL} ${BACK} accuracy {acc}"
+PY
+
+  cleanup
+  SERVER_PID=
+  sleep 1
+  PORT=$((PORT+1))
+done

.buildkite/test-pipeline.yaml

@@ -1373,4 +1373,22 @@ steps:
   num_gpus: 2
   working_dir: "/vllm-workspace"
   commands:
-  - bash .buildkite/scripts/scheduled_integration_test/qwen30b_a3b_fp8_block_ep_eplb.sh 0.8 200 8020 2 1
+  - bash .buildkite/scripts/scheduled_integration_test/qwen30b_a3b_fp8_block_ep_eplb.sh 0.8 200 8020 2 1
+
+- label: DeepSeek V2-Lite Async EPLB Accuracy
+  timeout_in_minutes: 60
+  gpu: h100
+  optional: true
+  num_gpus: 4
+  working_dir: "/vllm-workspace"
+  commands:
+  - bash .buildkite/scripts/scheduled_integration_test/deepseek_v2_lite_ep_async_eplb.sh 0.25 1319 8030
+
+- label: Qwen3-Next-80B-A3B-Instruct MTP Async EPLB Accuracy
+  timeout_in_minutes: 60
+  gpu: h100
+  optional: true
+  num_gpus: 4
+  working_dir: "/vllm-workspace"
+  commands:
+  - bash .buildkite/scripts/scheduled_integration_test/qwen3_next_mtp_async_eplb.sh 0.8 1319 8040

README.md

@@ -137,6 +137,7 @@ Compute Resources:
 - Alibaba Cloud
 - AMD
 - Anyscale
+- Arm
 - AWS
 - Crusoe Cloud
 - Databricks

docker/Dockerfile

@@ -150,8 +150,8 @@ ARG torch_cuda_arch_list='7.0 7.5 8.0 8.9 9.0 10.0 12.0'
 ENV TORCH_CUDA_ARCH_LIST=${torch_cuda_arch_list}
 #################### BASE BUILD IMAGE ####################
 
-#################### WHEEL BUILD IMAGE ####################
-FROM base AS build
+#################### CSRC BUILD IMAGE ####################
+FROM base AS csrc-build
 ARG TARGETPLATFORM
 
 ARG PIP_INDEX_URL UV_INDEX_URL
@@ -172,10 +172,13 @@ RUN --mount=type=cache,target=/root/.cache/uv \
     uv pip install --python /opt/venv/bin/python3 -r requirements/build.txt \
     --extra-index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.')
 
-COPY . .
-ARG GIT_REPO_CHECK=0
-RUN --mount=type=bind,source=.git,target=.git \
-    if [ "$GIT_REPO_CHECK" != "0" ]; then bash tools/check_repo.sh ; fi
+WORKDIR /workspace
+
+COPY pyproject.toml setup.py CMakeLists.txt ./
+COPY cmake cmake/
+COPY csrc csrc/
+COPY vllm/envs.py vllm/envs.py
+COPY vllm/__init__.py vllm/__init__.py
 
 # max jobs used by Ninja to build extensions
 ARG max_jobs=2
@@ -195,9 +198,11 @@ ARG SCCACHE_S3_NO_CREDENTIALS=0
 ARG VLLM_USE_PRECOMPILED=""
 ARG VLLM_MAIN_CUDA_VERSION=""
 
+# Use dummy version for csrc-build wheel (only .so files are extracted, version doesn't matter)
+ENV SETUPTOOLS_SCM_PRETEND_VERSION="0.0.0+csrc.build"
+
 # if USE_SCCACHE is set, use sccache to speed up compilation
 RUN --mount=type=cache,target=/root/.cache/uv \
-    --mount=type=bind,source=.git,target=.git \
     if [ "$USE_SCCACHE" = "1" ]; then \
         echo "Installing sccache..." \
         && curl -L -o sccache.tar.gz ${SCCACHE_DOWNLOAD_URL} \
@@ -223,7 +228,6 @@ ENV VLLM_TARGET_DEVICE=${vllm_target_device}
 ENV CCACHE_DIR=/root/.cache/ccache
 RUN --mount=type=cache,target=/root/.cache/ccache \
     --mount=type=cache,target=/root/.cache/uv \
-    --mount=type=bind,source=.git,target=.git  \
     if [ "$USE_SCCACHE" != "1" ]; then \
         # Clean any existing CMake artifacts
         rm -rf .deps && \
@@ -232,6 +236,52 @@ RUN --mount=type=cache,target=/root/.cache/ccache \
         export VLLM_DOCKER_BUILD_CONTEXT=1 && \
         python3 setup.py bdist_wheel --dist-dir=dist --py-limited-api=cp38; \
     fi
+#################### CSRC BUILD IMAGE ####################
+
+#################### WHEEL BUILD IMAGE ####################
+FROM base AS build
+ARG TARGETPLATFORM
+
+ARG PIP_INDEX_URL UV_INDEX_URL
+ARG PIP_EXTRA_INDEX_URL UV_EXTRA_INDEX_URL
+ARG PYTORCH_CUDA_INDEX_BASE_URL
+
+# install build dependencies
+COPY requirements/build.txt requirements/build.txt
+
+# This timeout (in seconds) is necessary when installing some dependencies via uv since it's likely to time out
+# Reference: https://github.com/astral-sh/uv/pull/1694
+ENV UV_HTTP_TIMEOUT=500
+ENV UV_INDEX_STRATEGY="unsafe-best-match"
+# Use copy mode to avoid hardlink failures with Docker cache mounts
+ENV UV_LINK_MODE=copy
+
+RUN --mount=type=cache,target=/root/.cache/uv \
+    uv pip install --python /opt/venv/bin/python3 -r requirements/build.txt \
+    --extra-index-url ${PYTORCH_CUDA_INDEX_BASE_URL}/cu$(echo $CUDA_VERSION | cut -d. -f1,2 | tr -d '.')
+
+WORKDIR /workspace
+
+COPY --from=csrc-build /workspace/dist /precompiled-wheels
+
+COPY . .
+
+ARG GIT_REPO_CHECK=0
+RUN --mount=type=bind,source=.git,target=.git \
+    if [ "$GIT_REPO_CHECK" != "0" ]; then bash tools/check_repo.sh ; fi
+
+ARG vllm_target_device="cuda"
+ENV VLLM_TARGET_DEVICE=${vllm_target_device}
+
+# Skip adding +precompiled suffix to version (preserves git-derived version)
+ENV VLLM_SKIP_PRECOMPILED_VERSION_SUFFIX=1
+
+RUN --mount=type=cache,target=/root/.cache/uv \
+    --mount=type=bind,source=.git,target=.git \
+    if [ "${vllm_target_device}" = "cuda" ]; then \
+        export VLLM_PRECOMPILED_WHEEL_LOCATION=$(ls /precompiled-wheels/*.whl); \
+    fi && \
+    python3 setup.py bdist_wheel --dist-dir=dist --py-limited-api=cp38
 
 # Install DeepGEMM from source
 ARG DEEPGEMM_GIT_REF

docs/assets/contributing/dockerfile-stages-dependency.png

@@ -18,6 +18,7 @@ Compute Resources:
 - Alibaba Cloud
 - AMD
 - Anyscale
+- Arm
 - AWS
 - Crusoe Cloud
 - Databricks

docs/community/sponsors.md

@@ -5,9 +5,6 @@ vLLM supports AMD GPUs with ROCm 6.3 or above, and torch 2.8.0 and above.
 !!! tip
     [Docker](#set-up-using-docker) is the recommended way to use vLLM on ROCm.
 
-!!! warning
-    There are no pre-built wheels for this device, so you must either use the pre-built Docker image or build vLLM from source.
-
 # --8<-- [end:installation]
 # --8<-- [start:requirements]
 

docs/getting_started/installation/gpu.rocm.inc.md

@@ -2,9 +2,6 @@
 
 vLLM initially supports basic model inference and serving on Intel GPU platform.
 
-!!! warning
-    There are no pre-built wheels for this device, so you need build vLLM from source. Or you can use pre-built images which are based on vLLM released versions.
-
 # --8<-- [end:installation]
 # --8<-- [start:requirements]