15:16 sarbes: I've had little time recently, but I'm pivoting towards the NIR route for my constant MUL MR (again). Covering all the different edge cases in PPIR is not fun.
15:17 sarbes: Same goes for all the other ops, which are not implemented atm.