Bitcount ?

Hi

I need an algoritm to do bitcount across an array, which ran in less than a cycle/byte on IA32/IA64.
I have written an algoritm for bit-parallel adition, and the best time wich I get is ~1.4 c/b.

I really appreciate your feedback.

Dimo Dimitrov

I'm not sure I've understood what exactly your problem is, but did you look at the popcnt instruction?

Dan