Clash of the Thread Pools
August 26, 2010 – 2:42 amRecently, I’ve been experimenting a little bit with different kinds of MT-safe containers. I wanted to compare a performance of various kinds of containers I’ve had lieing around. It turned out it’s more tricky than I expected, as under Windows, results vary wildly from one run to another. I’m not even speaking about container performance, just running same task can be many times slower (see this note for example). Include similar problems with managing the threads themselves (sometimes they wake up too late or not at all) and you should get the picture.








