[Libre-soc-dev] Coloquinte legalizer failure

Luke Kenneth Casson Leighton lkcl at lkcl.net
Sat Dec 5 02:07:18 GMT 2020

On 12/5/20, whygee at f-cpu.org <whygee at f-cpu.org> wrote:

> now, considering it's only for the first prototype(s),
> the real question is : is it practical/workable and
> which problems will we have to face, now AND later ?

targetting skywater 130nm PDK solves a huge number of issues.

FlexLib can be "tuned" (scaled) to a PDK, once you know the PDK.

however we need someone other that Staf to port FlexLib to non-NDA'd
PDKs such as Skywater 130nm, because aFoundry could claim that because
Staf developed FlexLib, he inadvertently or deliberately put NDA
details into the (example, Skywater 130nm) port, even though the port
is neither TSMC nor 180nm.

once we have *a* set of Cell Libraries that are not under NDA *then*
we can do dummy runs that approximate the "real" one.

another alternative is: we do not use FlexLib at all but instead use
alliance nsxlib 180nm which is crude, oversized, inefficient, but not
under NDA.  this would be a huge step backwards given that Staf's
library can fit 50% more gates per mm^2 than any other standard NDA'd
Foundry Cell Library.


More information about the Libre-soc-dev mailing list