I have run AFNI on a Xeon processor system before. It should compile and run fine. To get the best speed, you'll want to compile it yourself. You should use gcc 3.x, since that can produce faster code than gcc 2.x. You can then use some extra command line options to target the binary to the SSE instruction set on the more advanced Intel processors.
However, my experience with Xeon (3-4 years ago) was not very impressive. For the cost differential vs. Pentiums, at that time the speed improvement was marginal: 50% cost increase, 10% speed increase. What the situation is now, I have no idea.
bob cox