Skip to content

Allow inference test throughput to vary by 10% #2070

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
ko3n1g merged 1 commit into from
Oct 31, 2025
Merged

Allow inference test throughput to vary by 10% #2070

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
6 changes: 3 additions & 3 deletions tests/functional_tests/python_test_utils/test_inference_regular_pipeline.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -39,10 +39,10 @@ def test_inference_pipeline(golden_values_path: str, test_values_path: str) -> N
# First warmup iteration is excluded from throughput statistics.
throughput_sampled = median(output_current["throughput"][1:])

# 5 token/seconds is empirically observed to be within run variance.
# 10% is empirically observed to be within hardware variance.
assert (
throughput_sampled >= output_groundtruth["throughput"] - 5.0
), f"Throughput is slower than expected! Expected ~{output_groundtruth['throughput']} tok/s but benchmarked {output_current['throughput']} tok/s"
output_current["throughput"] >= 0.9 * output_groundtruth["throughput"]
), f"Throughput is slower than expected! Expected to be within 10% of ~{output_groundtruth['throughput']} tok/s but benchmarked {output_current['throughput']} tok/s"

# If throughput is significantly improved (> 20%), update golden values accordingly.
assert (
Expand Down