scatter3 is really slow

[Dr. Johannes Zellner]

hi,

scatter3 ist really slow here. It's MUCH slower than doing the same
plot in gnuplot.
What's the problem here?

Johannes
_______________________________________________
Help-octave mailing list
Help-octave@...
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave

[dbateman2]

Dr. Johannes Zellner wrote:
> hi,
>
> scatter3 ist really slow here. It's MUCH slower than doing the same
> plot in gnuplot.
> What's the problem here?
>
>
>  

Blame mathworks for considering that each point in a scatter point is a
separate patch object.. Then imagine the for loop that Octave needs to
have to send all that mess gnuplot.. I don't think there is much to do
here for the gnuplot backend though something might be possible with
other backends.

D.



--
David Bateman                                dbateman@...
35 rue Gambetta                              [phone] (Home)
92100 Boulogne-Billancourt FRANCE            [phone] (Mob)

_______________________________________________
Help-octave mailing list
Help-octave@...
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave

[Jaroslav Hajek-2]

On Thu, Jul 2, 2009 at 9:25 PM, David Bateman<dbateman@...> wrote:

> Dr. Johannes Zellner wrote:
>> hi,
>>
>> scatter3 ist really slow here. It's MUCH slower than doing the same
>> plot in gnuplot.
>> What's the problem here?
>>
>>
>>
>
> Blame mathworks for considering that each point in a scatter point is a
> separate patch object.. Then imagine the for loop that Octave needs to
> have to send all that mess gnuplot.. I don't think there is much to do
> here for the gnuplot backend though something might be possible with
> other backends.
>
> D.
>

If it is slowing things down this much (and I confirm, scatter takes
ages for 1000 points), what about making the Matlab compatible
behavior only optional? I can hardly imagine what it's good for to
have a separate handle for each point, except maybe scatter plots of
just several points. Matlab compatibility is generally good, but
surely not when it renders a function close to unusable.


--
RNDr. Jaroslav Hajek
computing expert & GNU Octave developer
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz
_______________________________________________
Help-octave mailing list
Help-octave@...
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave

[dbateman2]

Jaroslav Hajek wrote:

> On Thu, Jul 2, 2009 at 9:25 PM, David Bateman<dbateman@...> wrote:
>  
>> Dr. Johannes Zellner wrote:
>>    
>>> hi,
>>>
>>> scatter3 ist really slow here. It's MUCH slower than doing the same
>>> plot in gnuplot.
>>> What's the problem here?
>>>
>>>
>>>
>>>      
>> Blame mathworks for considering that each point in a scatter point is a
>> separate patch object.. Then imagine the for loop that Octave needs to
>> have to send all that mess gnuplot.. I don't think there is much to do
>> here for the gnuplot backend though something might be possible with
>> other backends.
>>
>> D.
>>
>>    
>
> If it is slowing things down this much (and I confirm, scatter takes
> ages for 1000 points), what about making the Matlab compatible
> behavior only optional? I can hardly imagine what it's good for to
> have a separate handle for each point, except maybe scatter plots of
> just several points. Matlab compatibility is generally good, but
> surely not when it renders a function close to unusable.
>
>
>  

The way to get a now matlab compatible behavior is something like

plot3(x,y,z,'r.')

The scatter3 command however allows each point in the scatter to have a
different color and that is the source of the issue... I don't set a way
of sending data for a set of points with a seperate color for each point
to gnuplot..

D.


--
David Bateman                                dbateman@...
35 rue Gambetta                              [phone] (Home)
92100 Boulogne-Billancourt FRANCE            [phone] (Mob)

_______________________________________________
Help-octave mailing list
Help-octave@...
https://www-old.cae.wisc.edu/mailman/listinfo/help-octave