This is what my animation function (i.e. the one that gets called by
`FuncAnimation`) looks like:
import numpy as np
...
def mpl_animation_function(n):
print "animating timestep: ", n
if n > 0:
previous_relevant_patch_indices =
np.ravel(patch_indices_per_timestep[n-1])
for index in previous_relevant_patch_indices:
(patches[index]).set_visible(False)
relevant_patch_indices = np.ravel(patch_indices_per_timestep[n])
for index in relevant_patch_indices:
(patches[index]).set_visible(True)
return patches,
`patches` is a pre-generated list of patches (possibly large), that have
already been added to an `axes` instance.
This function is awfully time-consuming as the number of patches becomes
large.
One idea I had was to parallelize the `for` loop, but likely that won't work
because of issues with the `axes` instance being accessed and modified in
parallel -- so I am afraid of fruitlessly spending time there. Do I have any
other options, or is parallelization possible?
···
--
View this message in context: http://matplotlib.1069221.n5.nabble.com/What-are-my-options-for-speeding-up-a-custom-function-called-by-FuncAnimation-tp45562.html
Sent from the matplotlib - users mailing list archive at Nabble.com.
What are you plotting? How big is this list that the loops are taking appreciable amounts of time?!? Are we talking seconds here or ms?
Have you done enough profiling to know exactly which line in here are slow? I don’t quite understand the np.ravel calls.
You might do better either with one (or many?) collection artists.
You might also look into just updating the artists you have.
Without some context of what these patches are it is really hard to help (or even really understand why this is slow).
Tom
···
On Sat, May 16, 2015 at 6:44 PM bmer <bhmerchant@…287…> wrote:
This is what my animation function (i.e. the one that gets called by
FuncAnimation) looks like:
import numpy as np
...
def mpl_animation_function(n):
print "animating timestep: ", n
if n > 0:
previous_relevant_patch_indices =
np.ravel(patch_indices_per_timestep[n-1])
for index in previous_relevant_patch_indices:
(patches[index]).set_visible(False)
relevant_patch_indices = np.ravel(patch_indices_per_timestep[n])
for index in relevant_patch_indices:
(patches[index]).set_visible(True)
return patches,
patches is a pre-generated list of patches (possibly large), that have
already been added to an axes instance.
This function is awfully time-consuming as the number of patches becomes
large.
One idea I had was to parallelize the for loop, but likely that won’t work
because of issues with the axes instance being accessed and modified in
parallel – so I am afraid of fruitlessly spending time there. Do I have any
other options, or is parallelization possible?
–
View this message in context: http://matplotlib.1069221.n5.nabble.com/What-are-my-options-for-speeding-up-a-custom-function-called-by-FuncAnimation-tp45562.html
Sent from the matplotlib - users mailing list archive at Nabble.com.
One dashboard for servers and applications across Physical-Virtual-Cloud
Widest out-of-the-box monitoring support with 50+ applications
Performance metrics, stats and reports that give you Actionable Insights
Deep dive visibility with transaction tracing using APM Insight.
http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
Matplotlib-users mailing list
Matplotlib-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/matplotlib-users