Skip to content

[#7588][fix] fixed the kv cache size parsing in test_perf.py AD backend #8092

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

Merged
merged 1 commit into from
Oct 2, 2025
Merged

[#7588][fix] fixed the kv cache size parsing in test_perf.py AD backend #8092

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
17 changes: 13 additions & 4 deletions tests/integration/defs/perf/test_perf.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -274,7 +274,8 @@ def import_allowed_perf_config():
PerfMetricType.OUTPUT_TOKEN_TIME:
re.compile(r"Average time-per-output-token \[TPOT\] \(ms\):\s+([\d\.]+)"),
PerfMetricType.KV_CACHE_SIZE:
re.compile(r".*Allocated ([\d\.]+) GiB for max tokens in paged KV cache.*"),
re.compile(r".*(?:Allocated ([\d\.]+) GiB for max tokens in paged KV cache|"
r"Final KV cache size after resize: ([\d\.]+) GiB).*"),
}
DISAGG_SERVER_METRICS_LOG_QUERIES = {
PerfMetricType.DISAGG_SERVER_E2EL:
Expand Down Expand Up @@ -1608,9 +1609,17 @@ def get_perf_result(self, outputs: Dict[int, str]) -> float:
for line in outputs[cmd_idx].split("\n")
]
print_info(outputs[cmd_idx].split("\n"))
metric_values = [
float(match.group(1)) for match in regex_matches if match
]
metric_values = []
for match in regex_matches:
if match:
# Handle multiple capture groups - use the first non-None group
value = None
for i in range(1, len(match.groups()) + 1):
if match.group(i) is not None:
value = match.group(i)
break
if value is not None:
metric_values.append(float(value))

if len(metric_values) == 0:
if self._build_script == "trtllm-build" and metric.metric_type == PerfMetricType.ENGINE_SIZE:
Expand Down