@User, @User: A lot of the SkyWater PDK has to be viewed in the light of coming from being a very much in-house thing at Cypress Semiconductor. Like most in-house stuff, I would assume, they got it to work but then just worked around the little quirks and errors and problems and never really polished it off for public consumption, because it was not being publicly consumed. The model binning is one of those things---they got used to designing with specific W and L devices. They characterized the devices that they put on some test wafer. Whatever software they used to get the SPICE models did not do a good job of fitting all parameters across continuous width and length, and so they just worked around it. That's my take on it; of course, I could be completely wrong, as I am only on the receiving end of the data.

Oh, so that's why they agreed to this 😉