That seems inefficient compared to a conventional register system. On the Mill, if my example code were implemented with pick, the code would be twice the size and always execute whether the condition was true or false (all for keeping track of previous arguments). A register system would just branch and execute the code or not. The mill makes up for this with parallelization but if a register system were equally parallel, it seems it would have better performance no?
Reply To: Introduction to the Mill CPU Programming Model lookaside