Closed nbingham1 closed 4 years ago
That's correct behavior. You're supposed to say mk_exclhi(a,b). The "exclhi(a,b)" is viewed as something that the circuit has to guarantee on its own (e.g. for a dualrail code).
Isn't prsim supposed to issue a warning if the "exclhi" spec is violated?
No, they are simply stripped out for prsim, as it doesn't have any support for invariant assertions.
So what are exclhi and excllo connected to then? Is it just a TODO at the moment?
They are used by lvp. More generally, the entire spec section will be re-done in the next year.