Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Fix test duration time inaccurate issue #183

Merged
merged 1 commit into from
Nov 6, 2024
Merged

Fix test duration time inaccurate issue #183

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 3 additions & 0 deletions evals/benchmark/stresscli/locust/aistress.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -120,6 +120,7 @@ def bench_main(self):
"faqgenfixed",
"faqgenbench",
]
test_start_time = time.time()
try:
start_ts = time.perf_counter()
with self.client.post(
Expand Down Expand Up @@ -150,6 +151,7 @@ def bench_main(self):
"response_string": resp.text,
"first_token_latency": time.perf_counter() - start_ts,
"total_latency": time.perf_counter() - start_ts,
"test_start_time": test_start_time,
}
else:
first_token_ts = None
Expand Down Expand Up @@ -184,6 +186,7 @@ def bench_main(self):
"response_string": complete_response,
"first_token_latency": first_token_ts - start_ts,
"total_latency": end_ts - start_ts,
"test_start_time": test_start_time,
}
reqdata = bench_package.respStatics(self.environment, reqData, respData)
logging.debug(f"Request data collected {reqdata}")
Expand Down
12 changes: 6 additions & 6 deletions evals/benchmark/stresscli/locust/tokenresponse.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -26,9 +26,7 @@ def respStatics(environment, req, resp):
num_token_input_prompt = -1

num_token_output = len(
tokenizer.encode(
resp["response_string"].lstrip().encode("utf-8").decode("unicode_escape"), add_special_tokens=False
)
tokenizer.encode(resp["response_string"].encode("utf-8").decode("unicode_escape"), add_special_tokens=False)
)

return {
Expand All @@ -37,6 +35,7 @@ def respStatics(environment, req, resp):
"first_token": resp["first_token_latency"] * 1000,
"next_token": (resp["total_latency"] - resp["first_token_latency"]) / (num_token_output - 1) * 1000,
"total_latency": resp["total_latency"] * 1000,
"test_start_time": resp["test_start_time"],
}


Expand All @@ -47,7 +46,6 @@ def staticsOutput(environment, reqlist):
e2e_lat = []
tokens_input = 0
tokens_output = 0
duration = environment.runner.stats.last_request_timestamp - environment.runner.stats.start_time

if len(reqlist) == 0:
logging.debug(f"len(reqlist): {len(reqlist)}, skip printing")
Expand All @@ -60,6 +58,8 @@ def staticsOutput(environment, reqlist):
e2e_lat.append(req["total_latency"])
tokens_output += req["tokens_output"]
tokens_input += req["tokens_input"]
test_start_time = req["test_start_time"]
duration = environment.runner.stats.last_request_timestamp - test_start_time

# Statistics for success response data only
if tokens_output == 0:
Expand All @@ -70,8 +70,8 @@ def staticsOutput(environment, reqlist):
" Output Tokens: {}, RPS: {:.2f}, Input Tokens per Second: {:.2f}, Output Tokens per Second: {:.2f}"
)
e2e_msg = "End to End latency(ms), P50: {:.2f}, P90: {:.2f}, P99: {:.2f}, Avg: {:.2f}"
first_msg = "First token latency(ms), P50: {:.2f}, P90: {:.2f}, P99: {:.2f}, Avg: {:.2f}"
next_msg = "Next token latency(ms), P50: {:.2f}, P90: {:.2f}, P99: {:.2f}, Avg: {:.2f}"
first_msg = "Time to First Token-TTFT(ms), P50: {:.2f}, P90: {:.2f}, P99: {:.2f}, Avg: {:.2f}"
next_msg = "Time Per Output Token-TPOT(ms), P50: {:.2f}, P90: {:.2f}, P99: {:.2f}, Avg: {:.2f}"
average_msg = "Average token latency(ms) : {:.2f}"
console_logger.warning("\n=================Total statistics=====================")
if tokens_output == 0:
Expand Down