Hi all, I've sent this query to c.l.python without any answer yet. I wondered if anyone here can either inform me or correct me. Basically, I was wondering if SMT (hyperthreading) in a Xeon gives similar (bad) performance characteristics to SMP in CPU-bound python apps? I am hoping not, because my guess is that a context switch between hyperthreads is considerably cheaper than an intra-CPU switch. But seeing as my knowledge of this area is fairly limited, I could be entirely wrong... seb