Interestingly, running a CPU intensive task like calculating Pi to 30000 digits for 6 simultaneous threads, is faster on my Quad core running at 2.4GHz than it is on my i7 (8 core) running at 1.6GHz.
So, even though the Quad can only handle 4 threads simultaneously without switching context, it’s fully a second faster than the i7 which can calculate for all six threads without switching at all. Obviously however, if I use a semaphore to limit the number of threads into a critical area by the number of cores, the i7 comes up trumps because of this fact.
The Quad completes the first four and then the last two in chunks. That chunking never takes place for the i7 as it can handle up to 8 threads at once, which makes it faster in this scenario.
Kinda jogs things back into perspective after all this “more cores for everybody” thing started. I’m thinking I’ll attempt to find a better balance between clock speed and cores going forward.