Daniel wrote: > Although I would note that it's very possible to embed the previous type > of loop into methods or functions if you wish. I normally write > object-oriented though and have heard that the multiprocessing library can > only be used with functions because it has trouble passing the objects > (although maybe explicitly specifying self would take care of that problem, > I've never tried it).
I'm not an expert, but fwiw another thing to consider is that the python multiprocessing library has to do a trick to make each "thread", it actually makes them as a much more expensive system process. the trouble is a basic limitation/feature of python where everything in the interpreter has to run in the same processor. so multi-processing in python inside a smaller loop or with NumPy can be problematic; better to stick with high-level MP in python I think. there's a lot of FAQ and threads out there about the Python Global Interpreter Lock (GIL) issue which is the reason for that. Hamish _______________________________________________ grass-user mailing list [email protected] http://lists.osgeo.org/mailman/listinfo/grass-user
