Opinon about PARIS / MGEN CAD suite

[dick_freebird]

PARIS / MGEN

This CAD suite was originally an in-house tool set at a European semi
house (Atmel, if I recall) and was "released into the wild" some years
back.

Wondering if it has gotten any traction, whether anyone out there has
succeeded in taping out to any foundry with it, any ideas on how much
work developing PDKs might be, and so on. It looks from the docs like
a pretty solid layout / verification suite, which so far is one area that
the open-source EDA community has failed to make any progress on.

 

[rfsystem]

I think MGEN is a far reaching approach to module generators not just PyCells which derive from the PCell concept. But there where no commercial tool support. Everything has to be interfaced. The history is very long if you consider L from Silicon Compiler Inc. as starting.

And every time there are two issues.

1. Could be the DRC rule base be 100% generic
2. Is the effort justifying the reuse gain

My thought is that a clever system which learn what layouters doing while doing and making abstracts and suggesting, being corrected, guided by the layouter and finally being a bionic processor instead of a programming system. That does not happen in nearly 30 years. So if you program layout you being the <1% rare layouter community which could abstract there own grafic ideas, understand the physics of the circuit, there logic organization and could transform that in a program flow and organization.