matploltib use gpu ?

Hello, i have a question.

I'm in Python 3.5, and can i use matploltib with GPU ?, can i use OpenGL or
PyOpenCL implemnetation for matploltib ?

Thanks for advance.

···

--
View this message in context: http://matplotlib.1069221.n5.nabble.com/matploltib-use-gpu-tp46633.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

Not at this time. However, there is another python project that does:
vispy.org

VisPy does a fantastic job of offloading to GPU and it is quite
straight-forward to use. There have been discussions between the devs of
the two projects to figure out how we can plug one project into the other
(most likely have vispy be an optional backend of sorts), but that is still
awhile down the road. Check it out!

Cheers!
Ben Root

···

On Tue, Jan 12, 2016 at 2:52 AM, nice <sazearte at live.fr> wrote:

Hello, i have a question.

I'm in Python 3.5, and can i use matploltib with GPU ?, can i use OpenGL or
PyOpenCL implemnetation for matploltib ?

Thanks for advance.

--
View this message in context:
http://matplotlib.1069221.n5.nabble.com/matploltib-use-gpu-tp46633.html
Sent from the matplotlib - users mailing list archive at Nabble.com.
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20160115/fb659011/attachment.html&gt;

I have not tried it, but another library for that is glumpy:
https://glumpy.github.io/

2016-01-15 14:50 GMT+00:00 Benjamin Root <ben.v.root at gmail.com>:

Not at this time. However, there is another python project that does:
vispy.org

VisPy does a fantastic job of offloading to GPU and it is quite
straight-forward to use. There have been discussions between the devs of
the two projects to figure out how we can plug one project into the other
(most likely have vispy be an optional backend of sorts), but that is still
awhile down the road. Check it out!

Cheers!
Ben Root

Hello, i have a question.

I'm in Python 3.5, and can i use matploltib with GPU ?, can i use OpenGL
or
PyOpenCL implemnetation for matploltib ?

Thanks for advance.

--
View this message in context:
http://matplotlib.1069221.n5.nabble.com/matploltib-use-gpu-tp46633.html
Sent from the matplotlib - users mailing list archive at Nabble.com.
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20160115/58062aa4/attachment-0001.html&gt;

···

On Tue, Jan 12, 2016 at 2:52 AM, nice <sazearte at live.fr> wrote:

Yes, glumpy is a sister project of VisPy that has largely been assimilated
into VisPy, if I understand it all correctly.

···

On Fri, Jan 15, 2016 at 9:57 AM, Ilya Flyamer <flyamer at gmail.com> wrote:

I have not tried it, but another library for that is glumpy:
https://glumpy.github.io/

2016-01-15 14:50 GMT+00:00 Benjamin Root <ben.v.root at gmail.com>:

Not at this time. However, there is another python project that does:
vispy.org

VisPy does a fantastic job of offloading to GPU and it is quite
straight-forward to use. There have been discussions between the devs of
the two projects to figure out how we can plug one project into the other
(most likely have vispy be an optional backend of sorts), but that is still
awhile down the road. Check it out!

Cheers!
Ben Root

On Tue, Jan 12, 2016 at 2:52 AM, nice <sazearte at live.fr> wrote:

Hello, i have a question.

I'm in Python 3.5, and can i use matploltib with GPU ?, can i use OpenGL
or
PyOpenCL implemnetation for matploltib ?

Thanks for advance.

--
View this message in context:
http://matplotlib.1069221.n5.nabble.com/matploltib-use-gpu-tp46633.html
Sent from the matplotlib - users mailing list archive at Nabble.com.
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20160115/e0da2f01/attachment.html&gt;

The first two points in its FAQ address this question:
http://glumpy.readthedocs.org/en/latest/faq.html

2016-01-15 15:03 GMT+00:00 Benjamin Root <ben.v.root at gmail.com>:

Yes, glumpy is a sister project of VisPy that has largely been assimilated
into VisPy, if I understand it all correctly.

I have not tried it, but another library for that is glumpy:
https://glumpy.github.io/

2016-01-15 14:50 GMT+00:00 Benjamin Root <ben.v.root at gmail.com>:

Not at this time. However, there is another python project that does:
vispy.org

VisPy does a fantastic job of offloading to GPU and it is quite
straight-forward to use. There have been discussions between the devs of
the two projects to figure out how we can plug one project into the other
(most likely have vispy be an optional backend of sorts), but that is still
awhile down the road. Check it out!

Cheers!
Ben Root

Hello, i have a question.

I'm in Python 3.5, and can i use matploltib with GPU ?, can i use
OpenGL or
PyOpenCL implemnetation for matploltib ?

Thanks for advance.

--
View this message in context:
http://matplotlib.1069221.n5.nabble.com/matploltib-use-gpu-tp46633.html
Sent from the matplotlib - users mailing list archive at Nabble.com.
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20160115/2f64404d/attachment.html&gt;

···

On Fri, Jan 15, 2016 at 9:57 AM, Ilya Flyamer <flyamer at gmail.com> wrote:

On Tue, Jan 12, 2016 at 2:52 AM, nice <sazearte at live.fr> wrote:

Ah, good to know about that update. Thank you!

···

On Fri, Jan 15, 2016 at 10:05 AM, Ilya Flyamer <flyamer at gmail.com> wrote:

The first two points in its FAQ address this question:
Frequently Asked Questions — glumpy v1.x documentation

2016-01-15 15:03 GMT+00:00 Benjamin Root <ben.v.root at gmail.com>:

Yes, glumpy is a sister project of VisPy that has largely been
assimilated into VisPy, if I understand it all correctly.

On Fri, Jan 15, 2016 at 9:57 AM, Ilya Flyamer <flyamer at gmail.com> wrote:

I have not tried it, but another library for that is glumpy:
https://glumpy.github.io/

2016-01-15 14:50 GMT+00:00 Benjamin Root <ben.v.root at gmail.com>:

Not at this time. However, there is another python project that does:
vispy.org

VisPy does a fantastic job of offloading to GPU and it is quite
straight-forward to use. There have been discussions between the devs of
the two projects to figure out how we can plug one project into the other
(most likely have vispy be an optional backend of sorts), but that is still
awhile down the road. Check it out!

Cheers!
Ben Root

On Tue, Jan 12, 2016 at 2:52 AM, nice <sazearte at live.fr> wrote:

Hello, i have a question.

I'm in Python 3.5, and can i use matploltib with GPU ?, can i use
OpenGL or
PyOpenCL implemnetation for matploltib ?

Thanks for advance.

--
View this message in context:
http://matplotlib.1069221.n5.nabble.com/matploltib-use-gpu-tp46633.html
Sent from the matplotlib - users mailing list archive at Nabble.com.
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20160115/e7266173/attachment-0001.html&gt;

ok, i can't use gpu.

But in this case it matplotlib can run multiple processor core?

I have quiad core processor and matploltib use 25% max processor. Can i use
more ?

···

--
View this message in context: http://matplotlib.1069221.n5.nabble.com/matploltib-use-gpu-tp46633p46663.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

matplotlib is not intrinsically multi-threaded, so it only uses one core
(and main obstacle there is the GIL, though parts of the Agg rendering
could probably be multithreaded if someone wanted to invest the effort).

However, if you are generating many plots in a row, the problem is
"embarrassingly parallel", and it's quite straightforward to use the python
multiprocessing library for this.

Mike

···

On Sat, Jan 16, 2016 at 4:39 AM, nice <sazearte at live.fr> wrote:

ok, i can't use gpu.

But in this case it matplotlib can run multiple processor core?

I have quiad core processor and matploltib use 25% max processor. Can i use
more ?

--
View this message in context:
http://matplotlib.1069221.n5.nabble.com/matploltib-use-gpu-tp46633p46663.html
Sent from the matplotlib - users mailing list archive at Nabble.com.
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

--
Michael Droettboom
Continuum Analytics
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20160125/6627fa7f/attachment.html&gt;

I should also point out that I rarely find times when matplotlib itself is
a bottleneck. Often, it is some other portion of the processing pipeline
that is the problem. I would utilize runsnakerun and do some profiling to
find out exactly where the pain points are in your code. I will note that,
in my experience, when it is the plotting that is slowing things down,
Basemap is typically the problem.

···

On Mon, Jan 25, 2016 at 9:06 AM, Michael Droettboom < mdroettboom at continuum.io> wrote:

matplotlib is not intrinsically multi-threaded, so it only uses one core
(and main obstacle there is the GIL, though parts of the Agg rendering
could probably be multithreaded if someone wanted to invest the effort).

However, if you are generating many plots in a row, the problem is
"embarrassingly parallel", and it's quite straightforward to use the python
multiprocessing library for this.

Mike

On Sat, Jan 16, 2016 at 4:39 AM, nice <sazearte at live.fr> wrote:

ok, i can't use gpu.

But in this case it matplotlib can run multiple processor core?

I have quiad core processor and matploltib use 25% max processor. Can i
use
more ?

--
View this message in context:
http://matplotlib.1069221.n5.nabble.com/matploltib-use-gpu-tp46633p46663.html
Sent from the matplotlib - users mailing list archive at Nabble.com.
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

--
Michael Droettboom
Continuum Analytics

_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20160125/862a299e/attachment.html&gt;

My problem: I have about 10 million point to plot.

with this example, moving in the graph is very slow.
Is there a way to have an intelligent zoom which removes invisible items on
the eye?

import matplotlib
matplotlib.use('TkAgg')
import numpy
from numpy import arange, sin, pi
from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg,
NavigationToolbar2TkAgg
from matplotlib.figure import Figure
import tkinter as Tk

root = Tk.Tk()
root.wm_title("Embedding in TK")

f = Figure(figsize=(5, 4), dpi=100)
a = f.add_subplot(111)

t = arange(0.0, 10000000, 1)
s = arange(0.0, 10000000, 1)
a.plot(t, s)

canvas = FigureCanvasTkAgg(f, master=root)
canvas.show()
canvas.get_tk_widget().pack(side=Tk.TOP, fill=Tk.BOTH, expand=1)

toolbar = NavigationToolbar2TkAgg(canvas, root)
toolbar.update()
canvas._tkcanvas.pack(side=Tk.TOP, fill=Tk.BOTH, expand=1)

def _quit():
    root.quit()
    root.destroy()

button = Tk.Button(master=root, text='Exit', command=_quit)
button.pack(side=Tk.BOTTOM)

Tk.mainloop()

···

--
View this message in context: http://matplotlib.1069221.n5.nabble.com/matploltib-use-gpu-tp46633p46677.html
Sent from the matplotlib - users mailing list archive at Nabble.com.

with this example, moving in the graph is very slow. Is there a way to

have an intelligent zoom which removes invisible items on the eye?

We already do this kind of optimization for line plots as in your example,
so I'm surprised it's slow for you. We don't do anything like that for
scatter plots though.

Mike

···

On Mon, Jan 25, 2016 at 12:35 PM, nice <sazearte at live.fr> wrote:

My problem: I have about 10 million point to plot.

with this example, moving in the graph is very slow.
Is there a way to have an intelligent zoom which removes invisible items on
the eye?

import matplotlib
matplotlib.use('TkAgg')
import numpy
from numpy import arange, sin, pi
from matplotlib.backends.backend_tkagg import FigureCanvasTkAgg,
NavigationToolbar2TkAgg
from matplotlib.figure import Figure
import tkinter as Tk

root = Tk.Tk()
root.wm_title("Embedding in TK")

f = Figure(figsize=(5, 4), dpi=100)
a = f.add_subplot(111)

t = arange(0.0, 10000000, 1)
s = arange(0.0, 10000000, 1)
a.plot(t, s)

canvas = FigureCanvasTkAgg(f, master=root)
canvas.show()
canvas.get_tk_widget().pack(side=Tk.TOP, fill=Tk.BOTH, expand=1)

toolbar = NavigationToolbar2TkAgg(canvas, root)
toolbar.update()
canvas._tkcanvas.pack(side=Tk.TOP, fill=Tk.BOTH, expand=1)

def _quit():
    root.quit()
    root.destroy()

button = Tk.Button(master=root, text='Exit', command=_quit)
button.pack(side=Tk.BOTTOM)

Tk.mainloop()

--
View this message in context:
http://matplotlib.1069221.n5.nabble.com/matploltib-use-gpu-tp46633p46677.html
Sent from the matplotlib - users mailing list archive at Nabble.com.
_______________________________________________
Matplotlib-users mailing list
Matplotlib-users at python.org
Matplotlib-users Info Page

--
Michael Droettboom
Continuum Analytics
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/matplotlib-users/attachments/20160126/6128d071/attachment.html&gt;

Yes, python turn has a 30% cpu instead of 25%
It there's an improvement in performance!

Why is there a 30% block? what art the rasterized

···

--
View this message in context: http://matplotlib.1069221.n5.nabble.com/matploltib-use-gpu-tp46633p46698.html
Sent from the matplotlib - users mailing list archive at Nabble.com.