@tduncan / Nvidia-Optimization-Stack-VLLM Mirror · read-only / commits / 4d29e91be8
Fork 0

A high-throughput and memory-efficient inference and serving engine for LLMs

Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git

Synced Nov 16, 03:15 UTC

[Misc] sort torch profiler table by kernel timing (#11813)

Divakar Verma committed Jan 7, 2025 at 20:57 UTC 4d29e91be84d27ca313d657eee92c067439a4c23
1 file changed +1 -1
benchmarks/benchmark_latency.py
+1 -1
@@ -52,7 +52,7 @@ def main(args: argparse.Namespace):
52 llm.generate(dummy_prompts,
53 sampling_params=sampling_params,
54 use_tqdm=False)
55 - print(p.key_averages())
55 + print(p.key_averages().table(sort_by="self_cuda_time_total"))
56 else:
57 start_time = time.perf_counter()
58 llm.generate(dummy_prompts,