On 2017-07-03 20:47, Xristos Xristoou wrote:
i have create an image processing python function.
my system have 4 cores + 4 threads.
i want to use multiprocessing to speed up my function,but anytime to use
multiprocessing packages my function is not faster and is 1 minute slowly. any
idea why ?first time use multiprocessing packages.
main function :
if __name__ == '__main__':
in_path="C:/Users/username/Desktop/in.tif"
out_path="C:/Users/username/Desktop/out.tif"
myfun(in_path, out_path)
time=3.4 minutes
with multiprocessing map :
if __name__ == '__main__':
p = Pool(processes=4)
in_path="C:/Users/username/Desktop/in.tif"
out_path="C:/Users/username/Desktop/out.tif"
result = p.map(myfun(in_path,out_path))
time=4.4 minutes
if __name__ == '__main__':
pool = multiprocessing.Pool(4)
in_path="C:/Users/username/Desktop/in.tif"
out_path="C:/Users/username/Desktop/out.tif"
pool.apply_async(myfun, args=(in_path,out_path,))
pool.close()
pool.join()
time=4.5 minutes
It looks like you have one function performing one task. It's not
splitting the task into multiple parts that can be run in parallel on
multiple cores.
--
https://mail.python.org/mailman/listinfo/python-list
