It's still an experimental module... hopefully it'll be polished up for the pygame 1.9 release. We included it so Nicholas could use it in the test runner (so it's possible to run all the tests in 512 threads/processes for example). It's nice for widely available 2-8 core machines... and the soon to be here 100+ core machines. So if you use it, beware that the API may change... but if you stick with just tmap, it should be fine, since that's supposed to copy the map api. It should help speed up loading on the first time you load the images(cold cache) -- with a single core machine, since then the disk IO is threaded.
On Wed, Aug 13, 2008 at 2:41 AM, Brian Fisher <[EMAIL PROTECTED]> wrote: > pygame.threads.tmap !!! > > Using it doesn't speed up the load in any way that I can notice on my single > core machine - but that's still a very cool function, cause rolling your own > "load in threads" would be a bit difficult... so why isn't it documented > anywhere on pygame.org? > > On Tue, Aug 12, 2008 at 2:56 AM, René Dudfield <[EMAIL PROTECTED]> wrote: >> >> > threads to load. Pygame can use multiple cpus to decode images... to >> > speed up loading. >> >> >> > > >
