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

Replace time.time() to time.perf_counter() for benchmarking. (#6178)

Browse Source 
Signed-off-by: Lifu Huang <lifu.hlf@gmail.com>
This commit is contained in:
Lifu Huang
2025-05-11 14:32:49 -07:00
committed by GitHub
parent e9a47f4cb5
commit 6e2da51561
61 changed files with 158 additions and 158 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
benchmark/benchmark_batch/benchmark_batch.py
View File

@@ -81,7 +81,7 @@ def send_batch_request(endpoint, prompts, gen_tokens, request_id):
}
data = {"text": prompts, "sampling_params": sampling_params}
start_time = time.time()
start_time = time.perf_counter()
try:
response = requests.post(
endpoint.base_url + "/generate", json=data, timeout=3600
@@ -90,7 +90,7 @@ def send_batch_request(endpoint, prompts, gen_tokens, request_id):
error = response.json()
raise RuntimeError(f"Request {request_id} failed: {error}")
result = response.json()
elapsed_time = (time.time() - start_time) * 1000 # Convert to ms
elapsed_time = (time.perf_counter() - start_time) * 1000 # Convert to ms
avg_per_prompt = elapsed_time / len(prompts) if prompts else 0
return request_id, elapsed_time, avg_per_prompt, True, len(prompts)
except Exception as e:
@@ -104,7 +104,7 @@ def run_benchmark(endpoint, batched_prompts, batch_size, gen_tokens):
num_requests = len(batched_prompts)
# Record start time for total latency
benchmark_start_time = time.time()
benchmark_start_time = time.perf_counter()
for i, batch_prompts in enumerate(batched_prompts):
request_id = i + 1
@@ -119,7 +119,7 @@ def run_benchmark(endpoint, batched_prompts, batch_size, gen_tokens):
results.append(result)
# Calculate total latency
total_latency = (time.time() - benchmark_start_time) * 1000 # Convert to ms
total_latency = (time.perf_counter() - benchmark_start_time) * 1000 # Convert to ms
return results, total_latency