Concurrently generate many matplotlib figures

Asked Jan 21 '19 at 16:24

Active Jan 21 '19 at 16:24

Viewed 75 times

I want to generate a large number (N) of PDF plots using matpotlib. Each plot requires reading the contents of a variable x without modifying it. When N is too large the code takes ages to finish. Is it possible to speed things up using some flavor of concurrency (e.g., by sending the job to generate each PDF to a different thread)?

from pylab import *

N = 1000
x = linspace(0,100,10001)

for n in range(N):
    fig = figure()
    plot(x,sin(x/(n+1)))
    savefig(f'{n+1}.pdf')
    close(fig)

asked Jan 21 '19 at 16:24

Mathieu

With multiprocessing you can speed things up by a factor < k, where k is the number of cores of your computer. Is that what you're after? – ImportanceOfBeingErnest Jan 21 '19 at 16:50
Yes, but I'm not sure if reading the same variable (x) from different processes/threads is possible? – Mathieu Jan 21 '19 at 17:02

Concurrently generate many matplotlib figures

0 Answers0