EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

feat: add e2e latency (#704)

Browse Source
This commit is contained in:
zhyncs
2024-07-24 05:51:10 +10:00
committed by GitHub
parent 268684439b
commit fa7ccb3316
1 changed files with 21 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
python/sglang/bench_serving.py
View File

@@ -264,6 +264,8 @@ class BenchmarkMetrics:
median_itl_ms: float median_itl_ms: float
std_itl_ms: float std_itl_ms: float
p99_itl_ms: float p99_itl_ms: float
mean_e2e_latency_ms: float
median_e2e_latency_ms: float
default_sharegpt_path = "ShareGPT_V3_unfiltered_cleaned_split.json" default_sharegpt_path = "ShareGPT_V3_unfiltered_cleaned_split.json"
@@ -467,6 +469,7 @@ def calculate_metrics(
itls: List[float] = [] itls: List[float] = []
tpots: List[float] = [] tpots: List[float] = []
ttfts: List[float] = [] ttfts: List[float] = []
e2e_latencies: List[float] = []
for i in range(len(outputs)): for i in range(len(outputs)):
if outputs[i].success: if outputs[i].success:
output_len = outputs[i].output_len output_len = outputs[i].output_len
@@ -480,6 +483,9 @@ def calculate_metrics(
tpots.append((outputs[i].latency - outputs[i].ttft) / (output_len - 1)) tpots.append((outputs[i].latency - outputs[i].ttft) / (output_len - 1))
itls += outputs[i].itl itls += outputs[i].itl
ttfts.append(outputs[i].ttft) ttfts.append(outputs[i].ttft)
e2e_latencies.append(outputs[i].latency)
completed += 1 completed += 1
else: else:
output_lens.append(0) output_lens.append(0)
@@ -513,6 +519,8 @@ def calculate_metrics(
median_itl_ms=np.median(itls or 0) * 1000, median_itl_ms=np.median(itls or 0) * 1000,
std_itl_ms=np.std(itls or 0) * 1000, std_itl_ms=np.std(itls or 0) * 1000,
p99_itl_ms=np.percentile(itls or 0, 99) * 1000, p99_itl_ms=np.percentile(itls or 0, 99) * 1000,
mean_e2e_latency_ms=np.mean(e2e_latencies) * 1000,
median_e2e_latency_ms=np.median(e2e_latencies) * 1000,
) )
return metrics, output_lens return metrics, output_lens
@@ -611,6 +619,15 @@ async def benchmark(
"Output token throughput (tok/s):", metrics.output_throughput "Output token throughput (tok/s):", metrics.output_throughput
) )
) )
print("{s:{c}^{n}}".format(s="End-to-End Latency", n=50, c="-"))
print(
"{:<40} {:<10.2f}".format("Mean E2E Latency (ms):", metrics.mean_e2e_latency_ms)
)
print(
"{:<40} {:<10.2f}".format(
"Median E2E Latency (ms):", metrics.median_e2e_latency_ms
)
)
print("{s:{c}^{n}}".format(s="Time to First Token", n=50, c="-")) print("{s:{c}^{n}}".format(s="Time to First Token", n=50, c="-"))
print("{:<40} {:<10.2f}".format("Mean TTFT (ms):", metrics.mean_ttft_ms)) print("{:<40} {:<10.2f}".format("Mean TTFT (ms):", metrics.mean_ttft_ms))
print("{:<40} {:<10.2f}".format("Median TTFT (ms):", metrics.median_ttft_ms)) print("{:<40} {:<10.2f}".format("Median TTFT (ms):", metrics.median_ttft_ms))
@@ -639,6 +656,8 @@ async def benchmark(
"total_input": metrics.total_input, "total_input": metrics.total_input,
"total_output": metrics.total_output, "total_output": metrics.total_output,
"total_output_retokenized": metrics.total_output_retokenized, "total_output_retokenized": metrics.total_output_retokenized,
"mean_e2e_latency": metrics.mean_e2e_latency_ms,
"median_e2e_latency": metrics.median_e2e_latency_ms,
"median_ttft": metrics.median_ttft_ms, "median_ttft": metrics.median_ttft_ms,
"median_itl": metrics.median_itl_ms, "median_itl": metrics.median_itl_ms,
"output_token_throughput": metrics.output_throughput, "output_token_throughput": metrics.output_throughput,
@@ -693,6 +712,8 @@ async def benchmark(
"itls": [output.itl for output in outputs], "itls": [output.itl for output in outputs],
"generated_texts": [output.generated_text for output in outputs], "generated_texts": [output.generated_text for output in outputs],
"errors": [output.error for output in outputs], "errors": [output.error for output in outputs],
"mean_e2e_latency_ms": metrics.mean_e2e_latency_ms,
"median_e2e_latency_ms": metrics.median_e2e_latency_ms,
} }
return result return result