check imshow performance #8143

Activity

1. tacaswell

   added

   Difficulty: Easyhttps://matplotlib.org/devdocs/devel/contribute.html#good-first-issueshttps://matplotlib.org/devdocs/devel/contribute.html#good-first-issues

   new-contributor-friendly

   on Feb 24, 2017
2. tacaswell

   added this to the 2.1 (next point release)This milestone has been deleted milestone on Feb 24, 2017
3. souravsingh commented on Feb 27, 2017

   

   souravsingh

   on Feb 27, 2017

   Contributor

   More actions

   @tacaswell I am interested in working on the issue. How do I benchmark the changes?
4. tacaswell commented on Feb 28, 2017

   

   tacaswell

   on Feb 28, 2017

   MemberAuthor

   More actions

   Using something like %timeit is a good start.
5. eozcelik commented on Mar 2, 2017

   

   eozcelik

   on Mar 2, 2017

   More actions

   X = np.arange(100)
   X = X.reshape(5, 20)
   
   fig = plt.figure()
   ax1 = fig.add_subplot(311)
   print(timeit.timeit("ax1.imshow(X, interpolation='nearest')", "from __main__ import ax1, X", number=10000))
   

   @tacaswell I have used the above snippet to test time performance and & seems to perform better than * after second or third decimal place in timeit output

   Should I be trying to localise the result in AxesImage.make_image ?
6. tacaswell commented on Sep 24, 2017

   

   tacaswell

   on Sep 24, 2017

   MemberAuthor

   More actions

   Closing this as #8966 re-wrote this code path and should have decreased the memory footprint.
7. tacaswell

   closed this as completedon Sep 24, 2017