Hi, I was using this script in ORFS to find the minimum clock period for a given design but I noticed that the period value written to the SDC depends on the period in the original SDC file provided. Is this the correct script to find the minimum clock period or is there a different command?

The period will depend on placement & routing. If you give a different input period you will likely get a different layout with a different min period

This is just post-synth. Sorry, forgot to mention that.

@Matt Liberty Please let us know what you think after your trip. Enjoy 🙂

yosys/abc depends on the clock period as well so I guess it will be similar. You can ask what min clock period is for a given implementation, not what the minimum could be under varying input conditions (the autotuner is about the best you can do there)

yosys/abc depends on the clock period as well so I guess it will be similar

The netlists generated were identical though...

Perhaps your variation isn't large enough to trigger a change. I would have to see a test case. In general the answer stands though as there is no command to tell you what is conceivable possible.

The variation was 100us to 1000us in clock period. Each time it generated the same netlist but the min clock period was roughly proportional to the input clock period (19us, 38us, 57us, 76us, ..., 190us)

It is common to stop optimizing once the target value is met. If you set it to something slower then the tool will stop at that value.

How is the optimization target decided? Is there any documentation for the

sta::worst_slack_cmd

? I couldn't find anything in the OpenSTA help command.

You should not be calling swig wrappers in the sta namespace. The command is report_wns

You should not be calling swig wrappers in the sta namespace.

The command was in the

write_ref_sdc.tcl

file in ORFS.

The value was needed in a script rather than the text report. The command should be used in general and is what will be documented