> So how about this proposal, we include the full set for the sake of > consistency -- it's not like vmsltu is significant different from a > performance modeling or correctness standpoint. The set should be fairly > mechanical. > > Can everyone support that?
Works for me. -- Regards Robin
