Matt Liberty comments

Results 1724 comments of


                                            Matt Liberty

Make antenna checker understand global route format

In https://github.com/The-OpenROAD-Project/OpenROAD/pull/4673 I've accepted some degradation in sky130hs/aes that originates from a new detour leading to more antenna fixing. Once the antenna fixing is improved please see if this design...

Assume port direction is input if io_type is unspecified in LEF.

I am not in favor of this change as I think this is fixing a symptom rather than the root cause. The four legal values of dbIoType are handled in...

sta slow at reporting clock skew

I recall them reporting an issue with threading, was there a similar performance issue reported? Is there an issue or PR?

sta slow at reporting clock skew

That represents a possible solution by threading. I'm still surprised that the single threaded time spread is so large. In any case it is a useful example.

Accessing a destroyed database from python should result in an exception, not SEGFAULT

Adding checking onto every single API call would be prohibitive in runtime and hard to do in practice. I don't see any practical solution to this. I think this is...

Accessing a destroyed database from python should result in an exception, not SEGFAULT

I agree it is a good goal but I see no practical implementation. I'm open to suggestions.

Accessing a destroyed database from python should result in an exception, not SEGFAULT

Its generated by swig from odb/src/swig/...

Accessing a destroyed database from python should result in an exception, not SEGFAULT

Try ``` from openroad import Design import odb db = Design.createDetachedDb() odb.dbChip.create(db) db.destroy(db) db.getChip().getBlock().getNets() ```

Accessing a destroyed database from python should result in an exception, not SEGFAULT

@macd that would handle the db object itself but all the other object handles you might have would still remain invalid (eg a dbInst). Do you see a way to...

Accessing a destroyed database from python should result in an exception, not SEGFAULT

I expect it is the same in tcl