Mike ins cross

Hi @GTumbush, this is a minor tweak to your #320. For reasons that are not clear to me, the per_instance option to covergroups causes dsim generate duplicated results. I am working that out with them, in the meantime, this kludge should ensure both xrun and dsim regressions produce good results.

I've tweaked the c.liu covergroup as I noticed it was not binning x16..x31 and had coverage for imm=0 which is not correct.