Hi, I'm doing LVS on a ring oscillator to get some familiarity with the tools. I found that by default, netgen says the circuits don't match (with

Networks match locally but not globally. Probably connections are swapped.

). As described in the netgen reference, this seems to be a false positive - it goes away and I get

Result: Circuits match uniquely

when using

-full

/

symmetry full

. Is using

-full

a good way to resolve these symmetries, or is it better to break the symmetry in the schematic somehow?

The "-full" option is the way netgen has always worked in the past, and is definitely the one to use here. The reason for introducing "-fast" is that standard cell designs were taking ages and ages to run the symmetry breaking at the end, after doing local matching. The reason for making it the default (which breaks backwards-compatibility) is that if you have a standard cell design, it may take 8-12 hours just to discover that there is something wrong with the circuit. There are probably several good solutions to this, but I have not had the time to look into them. When I get around to working on it, the -full option should go away, and netgen should go back to working like it always did, except that the long delay for standard cell designs will be solved properly. Meanwhile, the rule of thumb is that if it is not a standard cell design, you should use "-full".

ok, thank you Tim!