Corner Analysis varies the parameters of all transistors simultaneously.
Monte Carlo analysis doesn't do so. Instead, it varies the parameters of nominally matched transistors according to a pre-determined (from measurements) gaussian distribution. This allows you to investigate the effect of mismatch ( at a given corner )
I think it's not possible to extract Monte carlo's gaussian distb from corner data
Worst case model indicate the maximum variation of the process parameter, can I believe that it obey some distribution, such as uniform or gauss? I see from a subject that say many fabs use +/-4sigmas , in their min/max parameters. Is that right?
If the foundry has no monte carlo model, how can I do monte carlo simulation under the data of worst case model?