xinyun/vllm
1
0
Fork 0
mirror of https://git.datalinker.icu/vllm-project/vllm.git synced 2026-08-12 14:41:21 +08:00
Code Issues Packages Projects Releases Wiki Activity

[Bench] Split serve.py:main into async/async versions (#22405)

Browse Source 
Signed-off-by: Linkun <github@lkchen.net>
This commit is contained in:
lkchen 2025-08-06 23:05:07 -07:00 committed by GitHub
parent 2a4c825523
commit 4d4297e8fe
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 53 additions and 49 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
vllm/benchmarks/serve.py
View File

@ -948,7 +948,10 @@ def add_cli_args(parser: argparse.ArgumentParser):
) )
def main(args: argparse.Namespace): def main(args: argparse.Namespace) -> dict[str, Any]:
return asyncio.run(main_async(args))
async def main_async(args: argparse.Namespace) -> dict[str, Any]:
print(args) print(args)
random.seed(args.seed) random.seed(args.seed)
np.random.seed(args.seed) np.random.seed(args.seed)
@ -1025,8 +1028,7 @@ def main(args: argparse.Namespace):
gc.collect() gc.collect()
gc.freeze() gc.freeze()
benchmark_result = asyncio.run( benchmark_result = await benchmark(
benchmark(
endpoint_type=args.endpoint_type, endpoint_type=args.endpoint_type,
api_url=api_url, api_url=api_url,
base_url=base_url, base_url=base_url,
@ -1052,10 +1054,9 @@ def main(args: argparse.Namespace):
ramp_up_start_rps=args.ramp_up_start_rps, ramp_up_start_rps=args.ramp_up_start_rps,
ramp_up_end_rps=args.ramp_up_end_rps, ramp_up_end_rps=args.ramp_up_end_rps,
ready_check_timeout_sec=args.ready_check_timeout_sec, ready_check_timeout_sec=args.ready_check_timeout_sec,
)) )
# Save config and results to json # Save config and results to json
if args.save_result or args.append_result:
result_json: dict[str, Any] = {} result_json: dict[str, Any] = {}
# Setup # Setup
@ -1108,6 +1109,7 @@ def main(args: argparse.Namespace):
del benchmark_result[field] del benchmark_result[field]
# Save to file # Save to file
if args.save_result or args.append_result:
base_model_id = model_id.split("/")[-1] base_model_id = model_id.split("/")[-1]
max_concurrency_str = (f"-concurrency{args.max_concurrency}" max_concurrency_str = (f"-concurrency{args.max_concurrency}"
if args.max_concurrency is not None else "") if args.max_concurrency is not None else "")
@ -1129,3 +1131,5 @@ def main(args: argparse.Namespace):
outfile.write("\n") outfile.write("\n")
json.dump(result_json, outfile) json.dump(result_json, outfile)
save_to_pytorch_benchmark_format(args, result_json, file_name) save_to_pytorch_benchmark_format(args, result_json, file_name)
return result_json