Small change in clock period leads to wide swing in CTS TNS

[jeffng-or]

Subject

Describe the bug

Found during frequency sweep of asap7 ibex, where a small change in clock period leads to a much wider swing in CTS TNS. @maliberty requested filing GH issue to see if we can drive down the range.

Clock Period	CTS TNS
1001.8 ps	-49915 ps
1009.3 ps	-18862 ps

The overall CTS WNS/TNS chart looks like:

Expected Behavior

Reduced variation in TNS

Environment

b3f8033f64550f9601c8453d2cf7c178d7902133 - orfs git commit tag

To Reproduce

"min" value tarball: https://drive.google.com/file/d/1eAX7XIo6DWB1Y2afe-tg0q_fdV3dVwfW/view?usp=drive_link
"max" value tarball: https://drive.google.com/file/d/1IFMWERyTzKmT9rTYGk26U1r9auk30BAY/view?usp=drive_link

Relevant log output

Screenshots

No response

Additional Context

No response

[maliberty]

This is part of bigger goal to set better metrics limits but tns is so noisy the limit will be very large. Driving this variability down will help us set a more reasonable bound.

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed in 21 days if no further activity occurs. Remove the Stale label or comment to keep it open.