Fixes the misuse/mixuse of time.time()/time.monotonic() (#3220)

Signed-off-by: Tao He <sighingnow@gmail.com> Co-authored-by: simon-mo <simon.mo@hey.com>
2024-03-16 02:25:43 +08:00
parent 03d37f2441
commit 14b8ae02e7
5 changed files with 7 additions and 8 deletions
--- a/vllm/engine/async_llm_engine.py
+++ b/vllm/engine/async_llm_engine.py
@@ -604,8 +604,7 @@ class AsyncLLMEngine:
            >>> ...
        """
        # Preprocess the request.
-        # This should not be used for logging, as it is monotonic time.
-        arrival_time = time.monotonic()
+        arrival_time = time.time()

        try:
            stream = await self.add_request(