To support AA, I don't understand the numbers. From what I see, it takes the new code 24 unofficial seconds and 20.66 cpu cycles to run the 100000 iterations, while the old code takes only 22 unofficial seconds and just .01 of a cycle more to run the 100000 iterations.
Am I mis-reading something? Or is it trading off cpu work for speed?
------------------
Mike------------------
Visit STCC:
stcchat.com