Matt Liberty
Matt Liberty
The difficulty is that slack at an endpoint is a single number not a set of numbers based on these categories afaik. Its isn't a matter of presentation.
I talked with Tom and his suggestion is to try generating paths reg2reg paths using ``` report_checks -from [all_registers] -to [all_registers] -endpoint_count [llength [all_registers]] -unique_paths_to_endpoint ``` which should give a...
The slack histogram is not the same as report_checks. It is a value per endpoint not a set of paths.
I would prefer IO to IO as feedthru generally means a metal only connection between pins.
@rovinski I disagree and find it inconsistent with LEF.
I think it is a nice to have though not a high priority. To be clear this is command level timing rather than flow step level timing so it is...
We could either report whenever a command takes more than X time or just decide which commands are interesting to always report. The former will make the output somewhat non-deterministic...
Would ``` global_place rtl_macro_placer repair_design repair_timing global_route detailed_route ``` suffice? I suspect those are the big ticket items and wouldn't add too much verbosity.
repair_timing does give detail with -verbose. This would just be adding an overall time message.
Not a great example as the command didn't complete so you wouldn't get any reporting at the end. Reporting usage during progress is not what was requested here.