>George,
>
>I'd say it's due to the fact that we are still working in p-code. And there must be more overhead involved in passing the arguments even if the function does a CPU optimal OR at it's core.
>
David,

I can't see how that would make a measurable difference. Of course, my knowledge in this area is rather old, and perhaps you're aware of things that I'm not.

In the old days (16+ years ago), the tokens represented the offset in a jump table which pointed to the various entry points of the functions. I can't see why or how this could make a difference. As for the parameters, you'd still be passing them to whatever function you where calling (mathematical or bitwise). Unless there's a difference to the method of parameter passing, I can't see where that would make a difference. That's why my gut reaction is that it's the math co-processor.

If I'm wrong about the above, I'd appreciate it if you'd update me. I don't have the time anymore to keep up with such things, and if I took the time my creditors would be banging at the door < rbg >.

I love discussions like this one.