[Hardware] [Intel] Enable Multiprocessing and tensor parallel in CPU backend and update documentation (#6125)

2024-07-27 04:50:10 +08:00
parent aa4867791e
commit 3bbb4936dc
14 changed files with 403 additions and 89 deletions
--- a/vllm/engine/async_llm_engine.py
+++ b/vllm/engine/async_llm_engine.py
@@ -410,8 +410,6 @@ class AsyncLLMEngine:
            from vllm.executor.tpu_executor import TPUExecutorAsync
            executor_class = TPUExecutorAsync
        elif engine_config.device_config.device_type == "cpu":
-            assert distributed_executor_backend is None, (
-                "Distributed execution is not supported with the CPU backend.")
            from vllm.executor.cpu_executor import CPUExecutorAsync
            executor_class = CPUExecutorAsync
        elif engine_config.device_config.device_type == "openvino":