updates for EST fixes#4271
Conversation
Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>
…ate/OpenROAD-flow-scripts into secure-est_check_delay_calc Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>
Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>
Signed-off-by: Eder Monteiro <emrmonteiro@precisioninno.com>
There was a problem hiding this comment.
Code Review
This pull request lowers the core utilization for the gcd-ccs design to address a DPL failure, updates the tools/OpenROAD subproject commit, and updates the regression baselines in rules-base.json. However, the updated baselines show a significant degradation in timing metrics (e.g., cts__timing__setup__tns and globalroute__timing__setup__tns). Please ensure that this degradation is documented with an explicit agreement or a tracking issue as required by the repository guidelines.
| }, | ||
| "cts__timing__setup__tns": { | ||
| "value": -773.0, | ||
| "value": -1260.0, |
There was a problem hiding this comment.
The timing metrics show a significant degradation (e.g., cts__timing__setup__tns degrading from -773.0 to -1260.0 and globalroute__timing__setup__tns from -771.0 to -1530.0). Per the repository guidelines, merging updates to regression baselines with significant degradation is acceptable only if there is an explicit agreement or tracking issue to address the underlying timing issue later. Please document this agreement or reference the tracking issue in the PR description.
References
- It is acceptable to merge updates to regression baselines that show significant degradation, provided there is an explicit agreement to address the underlying issue later.
There was a problem hiding this comment.
This large degradations are explained by the fixes on the estimate_parasitics command. For this design, which uses set_delay_calculator prima, wire delay was always zero due to a bug. Now we're properly estimating parasitics for it.
designs/asap7/gcd-ccs/rules-base.json updates: