For applications that aren't multi-threaded, you can
gain some advantage of locking affinity to a single core - especially if the system, other than that, is mostly idle. Windows tends to schedule threads in a "load balancing" kind of way; for multicore CPUs with separate cache per core (rather than the shared approach of core2), it's a bit more optimal to stay on one core and avoid cache reloads and such.
I don't know how much you can really gain by this, as I haven't done any hard tests - but (placebo or not
) some single-threaded things do seem slightly smoother/faster with affinity locking on my AMD64x2.