[Intel GPU] support ray as distributed executor backend for XPU. (#20659)

Signed-off-by: Kunshang Ji <kunshang.ji@intel.com>

Author: jikunshang

.buildkite/scripts/hardware_ci/run-xpu-test.sh

@@ -27,6 +27,8 @@ docker run \
     "${image_name}" \
     sh -c '
     VLLM_USE_V1=1 python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager
+    VLLM_USE_V1=1 python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager -tp 2 --distributed-executor-backend ray
+    VLLM_USE_V1=1 python3 examples/offline_inference/basic/generate.py --model facebook/opt-125m --block-size 64 --enforce-eager -tp 2 --distributed-executor-backend mp
     cd tests
     pytest -v -s v1/core
 '

vllm/executor/ray_distributed_executor.py

@@ -62,7 +62,7 @@ class RayDistributedExecutor(DistributedExecutorBase):
 
     def _init_executor(self) -> None:
         self.forward_dag: Optional[ray.dag.CompiledDAG] = None
-        if envs.VLLM_USE_V1 and not current_platform.is_xpu():
+        if envs.VLLM_USE_V1:
             # V1 uses SPMD worker and compiled DAG
             os.environ["VLLM_USE_RAY_SPMD_WORKER"] = "1"
             os.environ["VLLM_USE_RAY_COMPILED_DAG"] = "1"