I just got winuae set up om my P3-650 with Win98SE and I found the speed in RTG mode really quick. I did some test renders with Cinema4D and found the emulator was 4 times faster than my A4000 68060. Hard drive access on the Amiga side was also much quicker. Going back to the A4000 felt like going from the A4000 to an A1200 '030. And I was impressed with compatabilty: i even got AGA demos running as fast as my A4000. Napalm ran really well as did Amidoom (as good as A4000 with CV64)
But you say that Amithlon "screamed"; do you mean it was faster than winuae in rtg emulation? It sounds logical as I understand that amithlon doesn't run Windows so this would free up the processor to run the amiga emulation.
One other question: having tried both P96 and CGX4 on my A4000. I preferred CGX because the mouse pointer moved smoother, icon dragging was smoother and the whole thing felt more responsive on my A4000 CV64. The winuae emulation is so accurate thet it emulates the stuff i don't like about P96. Sooo.. is there a way to use CGX as instead of P96 in winuae?