[CI/Build] enable Intel XPU test flow with prebuilt image (#37447)

Signed-off-by: wendyliu235 <wenjun.liu@intel.com>
2026-03-27 09:16:04 +08:00
parent f73bcb1c51
commit d86060122a
4 changed files with 397 additions and 0 deletions
--- a/.buildkite/ci_config_intel.yaml
+++ b/.buildkite/ci_config_intel.yaml
@@ -0,0 +1,23 @@
+name: vllm_intel_ci
+job_dirs:
+  - ".buildkite/intel_jobs"
+run_all_patterns:
+  - "docker/Dockerfile"
+  - "CMakeLists.txt"
+  - "requirements/common.txt"
+  - "requirements/xpu.txt"
+  - "requirements/build.txt"
+  - "requirements/test.txt"
+  - "setup.py"
+  - "csrc/"
+  - "cmake/"
+run_all_exclude_patterns:
+  - "docker/Dockerfile."
+  - "csrc/cpu/"
+  - "csrc/rocm/"
+  - "cmake/hipify.py"
+  - "cmake/cpu_extension.cmake"
+registries: public.ecr.aws/q9t5s3a7
+repositories:
+  main: "vllm-ci-test-repo"
+  premerge: "vllm-ci-test-repo"
--- a/.buildkite/image_build/image_build_xpu.sh
+++ b/.buildkite/image_build/image_build_xpu.sh
@@ -0,0 +1,34 @@
+#!/bin/bash
+set -e
+
+if [[ $# -lt 3 ]]; then
+  echo "Usage: $0 <registry> <repo> <commit>"
+  exit 1
+fi
+
+REGISTRY=$1
+REPO=$2
+BUILDKITE_COMMIT=$3
+
+# authenticate with AWS ECR
+aws ecr-public get-login-password --region us-east-1 | docker login --username AWS --password-stdin "$REGISTRY"
+aws ecr get-login-password --region us-east-1 | docker login --username AWS --password-stdin 936637512419.dkr.ecr.us-east-1.amazonaws.com
+
+# skip build if image already exists
+if ! docker manifest inspect "$REGISTRY"/"$REPO":"$BUILDKITE_COMMIT"-xpu &> /dev/null; then
+  echo "Image not found, proceeding with build..."
+else
+  echo "Image found"
+  exit 0
+fi
+
+# build
+docker build \
+  --file docker/Dockerfile.xpu \
+  --build-arg max_jobs=16 \
+  --build-arg buildkite_commit="$BUILDKITE_COMMIT" \
+  --tag "$REGISTRY"/"$REPO":"$BUILDKITE_COMMIT"-xpu \
+  --progress plain .
+
+# push
+docker push "$REGISTRY"/"$REPO":"$BUILDKITE_COMMIT"-xpu
--- a/.buildkite/intel_jobs/test-intel.yaml
+++ b/.buildkite/intel_jobs/test-intel.yaml
@@ -0,0 +1,64 @@
+group: Intel
+steps:
+  - label: ":docker: Build XPU image"
+    soft_fail: true
+    depends_on: []
+    key: image-build-xpu
+    commands:
+      - bash -lc '.buildkite/image_build/image_build_xpu.sh "public.ecr.aws/q9t5s3a7" "vllm-ci-test-repo" "$BUILDKITE_COMMIT"'
+    env:
+      DOCKER_BUILDKIT: "1"
+    retry:
+      automatic:
+        - exit_status: -1  # Agent was lost
+          limit: 2
+        - exit_status: -10  # Agent was lost
+          limit: 2
+  - label: "XPU example Test"
+    depends_on:
+      - image-build-xpu
+    timeout_in_minutes: 30
+    device: intel_gpu
+    no_plugin: true
+    env:
+      REGISTRY: "public.ecr.aws/q9t5s3a7"
+      REPO: "vllm-ci-test-repo"
+    source_file_dependencies:
+      - vllm/
+      - .buildkite/intel_jobs/test-intel.yaml 
+    commands:
+      - >-
+        bash .buildkite/scripts/hardware_ci/run-intel-test.sh
+        'pip install tblib==3.1.0 &&
+        python3 examples/basic/offline_inference/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager &&
+        python3 examples/basic/offline_inference/generate.py --model facebook/opt-125m --block-size 64 -O3 -cc.cudagraph_mode=NONE &&
+        python3 examples/basic/offline_inference/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager -tp 2 --distributed-executor-backend mp &&
+        python3 examples/basic/offline_inference/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager --attention-backend=TRITON_ATTN &&
+        python3 examples/basic/offline_inference/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager --quantization fp8 &&
+        python3 examples/basic/offline_inference/generate.py --model superjob/Qwen3-4B-Instruct-2507-GPTQ-Int4 --block-size 64 --enforce-eager --max-model-len 8192 &&
+        python3 examples/basic/offline_inference/generate.py --model ibm-research/PowerMoE-3b --block-size 64 --enforce-eager -tp 2 &&
+        python3 examples/basic/offline_inference/generate.py --model ibm-research/PowerMoE-3b --block-size 64 --enforce-eager -tp 2 --enable-expert-parallel'
+  - label: "XPU V1 test"
+    depends_on:
+      - image-build-xpu
+    timeout_in_minutes: 30
+    device: intel_gpu
+    no_plugin: true
+    env:
+      REGISTRY: "public.ecr.aws/q9t5s3a7"
+      REPO: "vllm-ci-test-repo"
+    source_file_dependencies:
+      - vllm/
+      - .buildkite/intel_jobs/test-intel.yaml 
+    commands:
+      - >-
+        bash .buildkite/scripts/hardware_ci/run-intel-test.sh
+        'cd tests &&
+        pytest -v -s v1/core --ignore=v1/core/test_reset_prefix_cache_e2e.py --ignore=v1/core/test_scheduler_e2e.py &&
+        pytest -v -s v1/engine --ignore=v1/engine/test_output_processor.py &&
+        pytest -v -s v1/sample --ignore=v1/sample/test_logprobs.py --ignore=v1/sample/test_logprobs_e2e.py &&
+        pytest -v -s v1/worker --ignore=v1/worker/test_gpu_model_runner.py --ignore=v1/worker/test_worker_memory_snapshot.py &&
+        pytest -v -s v1/structured_output &&
+        pytest -v -s v1/test_serial_utils.py &&
+        pytest -v -s v1/spec_decode --ignore=v1/spec_decode/test_max_len.py --ignore=v1/spec_decode/test_tree_attention.py --ignore=v1/spec_decode/test_speculators_eagle3.py --ignore=v1/spec_decode/test_acceptance_length.py &&
+        pytest -v -s v1/kv_connector/unit --ignore=v1/kv_connector/unit/test_multi_connector.py --ignore=v1/kv_connector/unit/test_nixl_connector.py --ignore=v1/kv_connector/unit/test_example_connector.py --ignore=v1/kv_connector/unit/test_lmcache_integration.py'
--- a/.buildkite/scripts/hardware_ci/run-intel-test.sh
+++ b/.buildkite/scripts/hardware_ci/run-intel-test.sh
@@ -0,0 +1,276 @@
+#!/bin/bash
+
+# This script runs tests inside the Intel XPU docker container.
+# It mirrors the structure of run-amd-test.sh while keeping Intel-specific
+# container setup and allowing commands to be sourced from YAML or env.
+#
+# Command sources (in priority order):
+#   1) VLLM_TEST_COMMANDS env var (preferred, preserves quoting)
+#   2) Positional args (legacy)
+#   3) One or more YAML files with a commands list (test-area style)
+###############################################################################
+set -o pipefail
+
+DRY_RUN=${DRY_RUN:-0}
+if [[ "${1:-}" == "--dry-run" ]]; then
+  DRY_RUN=1
+  shift
+fi
+
+# Export Python path
+export PYTHONPATH=".."
+
+###############################################################################
+# Helper Functions
+###############################################################################
+
+cleanup_docker() {
+  docker_root=$(docker info -f '{{.DockerRootDir}}')
+  if [ -z "$docker_root" ]; then
+    echo "Failed to determine Docker root directory." >&2
+    exit 1
+  fi
+  echo "Docker root directory: $docker_root"
+
+  disk_usage=$(df "$docker_root" | tail -1 | awk '{print $5}' | sed 's/%//')
+  threshold=70
+  if [ "$disk_usage" -gt "$threshold" ]; then
+    echo "Disk usage is above $threshold%. Cleaning up Docker images and volumes..."
+    docker image prune -f
+    docker volume prune -f && docker system prune --force --filter "until=72h" --all
+    echo "Docker images and volumes cleanup completed."
+  else
+    echo "Disk usage is below $threshold%. No cleanup needed."
+  fi
+}
+
+re_quote_pytest_markers() {
+  local input="$1"
+  local output=""
+  local collecting=false
+  local marker_buf=""
+
+  local flat="${input//$'\n'/ }"
+  local restore_glob
+  restore_glob="$(shopt -p -o noglob 2>/dev/null || true)"
+  set -o noglob
+  local -a words
+  read -ra words <<< "$flat"
+  eval "$restore_glob"
+
+  for word in "${words[@]}"; do
+    if $collecting; then
+      if [[ "$word" == *"'"* ]]; then
+        if [[ -n "$marker_buf" ]]; then
+          output+="${marker_buf} "
+          marker_buf=""
+        fi
+        output+="${word} "
+        collecting=false
+        continue
+      fi
+
+      local is_boundary=false
+      case "$word" in
+        "&&"|"||"|";"|"|")
+          is_boundary=true ;;
+        --*)
+          is_boundary=true ;;
+        -[a-zA-Z])
+          is_boundary=true ;;
+        */*)
+          is_boundary=true ;;
+        *.py|*.py::*)
+          is_boundary=true ;;
+        *=*)
+          if [[ "$word" =~ ^[A-Z_][A-Z0-9_]*= ]]; then
+            is_boundary=true
+          fi
+          ;;
+      esac
+
+      if $is_boundary; then
+        if [[ "$marker_buf" == *" "* || "$marker_buf" == *"("* ]]; then
+          output+="'${marker_buf}' "
+        else
+          output+="${marker_buf} "
+        fi
+        collecting=false
+        marker_buf=""
+        if [[ "$word" == "-m" || "$word" == "-k" ]]; then
+          output+="${word} "
+          collecting=true
+        else
+          output+="${word} "
+        fi
+      else
+        if [[ -n "$marker_buf" ]]; then
+          marker_buf+=" ${word}"
+        else
+          marker_buf="${word}"
+        fi
+      fi
+    elif [[ "$word" == "-m" || "$word" == "-k" ]]; then
+      output+="${word} "
+      collecting=true
+      marker_buf=""
+    else
+      output+="${word} "
+    fi
+  done
+
+  if $collecting && [[ -n "$marker_buf" ]]; then
+    if [[ "$marker_buf" == *" "* || "$marker_buf" == *"("* ]]; then
+      output+="'${marker_buf}'"
+    else
+      output+="${marker_buf}"
+    fi
+  fi
+
+  echo "${output% }"
+}
+
+apply_intel_test_overrides() {
+  local cmds="$1"
+  # Placeholder for Intel-specific exclusions/overrides.
+  echo "$cmds"
+}
+
+is_yaml_file() {
+  local p="$1"
+  [[ -f "$p" && "$p" == *.yaml ]]
+}
+
+extract_yaml_commands() {
+  local yaml_path="$1"
+  awk '
+    $1 == "commands:" { in_cmds=1; next }
+    in_cmds && $0 ~ /^[[:space:]]*-[[:space:]]/ {
+      sub(/^[[:space:]]*-[[:space:]]/, "");
+      print;
+      next
+    }
+    in_cmds && $0 ~ /^[^[:space:]]/ { exit }
+  ' "$yaml_path"
+}
+
+###############################################################################
+# Main
+###############################################################################
+
+default_image_name="${REGISTRY}/${REPO}:${BUILDKITE_COMMIT}-xpu"
+#default_image_name="public.ecr.aws/q9t5s3a7/vllm-ci-test-repo:${BUILDKITE_COMMIT}-xpu"
+image_name="${IMAGE_TAG_XPU:-${default_image_name}}"
+container_name="xpu_${BUILDKITE_COMMIT}_$(tr -dc A-Za-z0-9 < /dev/urandom | head -c 10; echo)"
+
+# ---- Command source selection ----
+commands=""
+if [[ -n "${VLLM_TEST_COMMANDS:-}" ]]; then
+  commands="${VLLM_TEST_COMMANDS}"
+  echo "Commands sourced from VLLM_TEST_COMMANDS (quoting preserved)"
+elif [[ $# -gt 0 ]]; then
+  all_yaml=true
+  for arg in "$@"; do
+    if ! is_yaml_file "$arg"; then
+      all_yaml=false
+      break
+    fi
+  done
+
+  if $all_yaml; then
+    for yaml in "$@"; do
+      mapfile -t COMMANDS < <(extract_yaml_commands "$yaml")
+      if [[ ${#COMMANDS[@]} -eq 0 ]]; then
+        echo "Error: No commands found in ${yaml}" >&2
+        exit 1
+      fi
+      for cmd in "${COMMANDS[@]}"; do
+        if [[ -z "$commands" ]]; then
+          commands="${cmd}"
+        else
+          commands+=" && ${cmd}"
+        fi
+      done
+    done
+    echo "Commands sourced from YAML files: $*"
+  else
+    commands="$*"
+    echo "Commands sourced from positional args (legacy mode)"
+  fi
+else
+  SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+  DEFAULT_YAML="${SCRIPT_DIR}/intel-test.yaml"
+  if [[ ! -f "${DEFAULT_YAML}" ]]; then
+    echo "Error: YAML file not found: ${DEFAULT_YAML}" >&2
+    exit 1
+  fi
+  mapfile -t COMMANDS < <(extract_yaml_commands "${DEFAULT_YAML}")
+  if [[ ${#COMMANDS[@]} -eq 0 ]]; then
+    echo "Error: No commands found in ${DEFAULT_YAML}" >&2
+    exit 1
+  fi
+  for cmd in "${COMMANDS[@]}"; do
+    if [[ -z "$commands" ]]; then
+      commands="${cmd}"
+    else
+      commands+=" && ${cmd}"
+    fi
+  done
+  echo "Commands sourced from default YAML: ${DEFAULT_YAML}"
+fi
+
+if [[ -z "$commands" ]]; then
+  echo "Error: No test commands provided." >&2
+  exit 1
+fi
+
+echo "Raw commands: $commands"
+commands=$(re_quote_pytest_markers "$commands")
+echo "After re-quoting: $commands"
+commands=$(apply_intel_test_overrides "$commands")
+echo "Final commands: $commands"
+
+# Dry-run mode prints final commands and exits before Docker.
+if [[ "$DRY_RUN" == "1" ]]; then
+  echo "DRY_RUN=1 set, skipping Docker execution."
+  exit 0
+fi
+
+# --- Docker housekeeping ---
+cleanup_docker
+
+# --- Build or pull test image ---
+if [[ -n "${IMAGE_TAG_XPU:-}" ]]; then
+  echo "Using prebuilt XPU image: ${IMAGE_TAG_XPU}"
+  docker pull "${IMAGE_TAG_XPU}"
+else
+  echo "Using prebuilt XPU image: ${image_name}"
+  docker pull "${image_name}"
+fi
+
+remove_docker_container() {
+  docker rm -f "${container_name}" || true
+  docker image rm -f "${image_name}" || true
+  docker system prune -f || true
+}
+trap remove_docker_container EXIT
+
+# --- Single-node job ---
+
+if [[ -z "${ZE_AFFINITY_MASK:-}" ]]; then
+  echo "Warning: ZE_AFFINITY_MASK is not set. Proceeding without device affinity." >&2
+fi
+
+docker run \
+    --device /dev/dri:/dev/dri \
+    --net=host \
+    --ipc=host \
+    --privileged \
+    -v /dev/dri/by-path:/dev/dri/by-path \
+    --entrypoint="" \
+    -e "HF_TOKEN=${HF_TOKEN:-}" \
+    -e "ZE_AFFINITY_MASK=${ZE_AFFINITY_MASK:-}" \
+    -e "CMDS=${commands}" \
+    --name "${container_name}" \
+    "${image_name}" \
+    bash -c 'set -e; echo "ZE_AFFINITY_MASK is ${ZE_AFFINITY_MASK:-}"; eval "$CMDS"'