Yes, and we won't be able to conclude anything. This is micro-optimization which doesn't give meaningful results.
That's why I said "if what you say is true", i.e. if the claimed slowdown from the extra argument decoding step actually exists. There are two possibilities: (1) It really does slow down the execution of all bytescodes. In that case, move it into the handler code for the relevant bytecodes, where it won't affect the speed of anything else. (2) Otherwise, leave it where it is. Either way, I don't see any justification to reject the idea of having 2-arg opcodes because of speed, which is what the original post seemed to be saying. Or maybe I misunderstood. Greg Ewing, Computer Science Dept, +--------------------------------------+ University of Canterbury, | A citizen of NewZealandCorp, a | Christchurch, New Zealand | wholly-owned subsidiary of USA Inc. | greg@cosc.canterbury.ac.nz +--------------------------------------+