I've had to edit the simulate_and_measure timeout ...
# chipaloozad
diadatp
03/24/2024, 5:16 AMI've had to edit the simulate_and_measure timeout (https://github.com/efabless/cace/blob/743301ab6d719cd8e1f6af5ed314386dd3f340a5/cace/common/cace_launch.py#L320) as the INL and DNL simulations take a very long time to complete. Is there a workaround where I can return results form a single spice run that does the alter loop and then pass it along to the measure script?
diadatp
03/24/2024, 6:02 AMThe documentation mentions {cond|minimum} {cond|maximum} {cond|stepsize} {cond|steps}.
t
Tim Edwards
03/24/2024, 1:03 PMThere should be a way to do that, but I should also probably have a way to override that 300 second limit. It's very hard to account for problems arising from simulations which have become stuck vs. simulations which are simply running long. A "settings" option in CACE to set the timeout, though, should be easy enough to implement.
Tim Edwards
03/24/2024, 1:48 PMIf I did a quick update to CACE and added an override for that 300 second timeout, would that solve your immediate problem?
Tim Edwards
03/24/2024, 9:22 PMAfter thinking about it for a short while, I realized that the point of the timeout was to keep the simulations from hanging CACE. But as it stands, the timeout causes CACE to hang, so it's less than useless. I removed the timeout, so if you update CACE, you should no longer have the timeout issue.
@Robin Tsang also encountered this problem, so an update to CACE should work for both of you.
d
diadatp
03/25/2024, 8:26 AMUpdated. Thank you for the quick update.
5 Views