Augusto Berndt
Augusto Berndt
Currently we do not overlap the routability and timing-driven overflow thresholds to avoid this and other potential issues on the interactions between the two modes.
I ran a secure-CI in: https://jenkins.openroad.tools/blue/organizations/jenkins/OpenROAD-flow-scripts-Private/detail/secure-gpl-use-unscaled-overflow/1/pipeline
Hi! Recently we have noticed divergences happening on stage 3-1 (skip io), and we made modifications to gpl regarding a bivariate normal distribution adjustment made with macros, reducing its effect...
Hi @mikesinouye, what stage exactly do you have the error? is it on skip io stage 3-1, or during global placement itself on stage 3-3?
Indeed, that's a big jump after the last timing-driven iteration. I would have to take a look on the debug mode, it is unfortunate it is a private PDK. I...
Concerning mega boom, I am investigating the issue and should provide new insights soon enough.
Hello @oharboe. I apologize, but we have not yet been able to determine the exact cause of the divergence. There is evidence suggesting it may be related to how we...
> @gudeh Could it be that this problem has been there all along and that it is just bad luck that some initial conditions for the flow has changed and...
Allow me to share some thoughts. From the experiments I tried, there was two that lead to a convergence: - Increasing the bin size by 10x or 5x, As of...
Hi. I implemented the idea of saving a snapshot and finishing with it instead of throwing an error. For this megaboom, it finished with the following placement. Which looks a...