Daniel J Sebald wrote: > produce similar results. The bit-shift/bit-or version, not only being > rather ugly, is slower than the multiply/add version by a factor of two. Eh, I was wrong on that; had a variable name wrong. They are roughly the same time. Dan