Kent Johnson, 27.11.2009 14:38: > On Fri, Nov 27, 2009 at 4:57 AM, OkaMthembo wrote: >> Is there a python implementation that takes advantage of all cores on modern >> multicore machines? > > Presumably you mean something like, "Is there a python implementation > that can run multiple compute-bound processes on multiple cores > concurrently." > > Some options: > - the multiprocessing module in the std lib - here is an example of > using it with numpy: > http://folk.uio.no/sturlamo/python/multiprocessing-tutorial.pdf
In case this is about some kind of numeric computation that doesn't require Python object types, another option is to use Cython and put the compute intensive code into a "nogil" function that you can run multi-threaded. If the above doesn't apply, it may be quite possible that the GIL is not even the limiting factor. Unless the OP provides more insight, this is impossible to answer. Stefan _______________________________________________ Tutor maillist - Tutor@python.org To unsubscribe or change subscription options: http://mail.python.org/mailman/listinfo/tutor
