You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The lines that converted each closed collection of points to a numpy array are repeated in paths.Path which is where the data is sent to next. I removed these lines and made sure that the appropriate zeros are still added. This a pretty significant speed-up in some cases. I was getting a 10% speed-up in overall time for using plot_surface with lots of points.
SourceForge Comments
On Wed Dec 22 01:52:25 2010, notmuchtotell wrote:
The new patch I've attached will work with both numpy and masked numpy arrays. The speed up isn't quite as big, but it is still substantial.
On Tue Dec 21 11:06:29 2010, mdboom wrote:
Does this break if xy is a Numpy array or masked array to begin with? We can special case this for the faster performance on lists, but it needs to work with arrays as well.
SourceForge History
On Wed Dec 22 01:51:29 2010, by notmuchtotell: File Added: 396688: set_verts_cleanup_2.patch
On Wed Dec 22 01:47:47 2010, by notmuchtotell: File Deleted: 396374:
On Fri Dec 17 12:10:35 2010, by notmuchtotell: File Added: 396374: set_verts_cleanup.patch
The text was updated successfully, but these errors were encountered:
Either way, sadly we cannot get the original patch from SourceForge so this issue should be closed.
If the original author (notmuchtotell) ever finds this, and has the original patch, and it hasn't been merged. Please let us know and we can integrate it.
Original report at SourceForge, opened Fri Dec 17 12:10:35 2010
The lines that converted each closed collection of points to a numpy array are repeated in paths.Path which is where the data is sent to next. I removed these lines and made sure that the appropriate zeros are still added. This a pretty significant speed-up in some cases. I was getting a 10% speed-up in overall time for using plot_surface with lots of points.
SourceForge Comments
On Wed Dec 22 01:52:25 2010, notmuchtotell wrote:
The new patch I've attached will work with both numpy and masked numpy arrays. The speed up isn't quite as big, but it is still substantial.
On Tue Dec 21 11:06:29 2010, mdboom wrote:
Does this break if xy is a Numpy array or masked array to begin with? We can special case this for the faster performance on lists, but it needs to work with arrays as well.
SourceForge History
The text was updated successfully, but these errors were encountered: