visualpython-users

[Visualpython-users] visual 5.0 speed ?

[Gary P. <gar...@gm...>]

I've finally upgraded to 5.0 on my XP machine.

I have a subjective sense that it might be slower than 3.0.    Before
I try to reinstall 3.0 to check, I thought I'd ask ...

What's known about the speed of 5.0 vs 3.0 ?

-Gary

Re: [Visualpython-users] visual 5.0 speed ?

[Bruce S. <Bru...@nc...>]

Lots of tests I've run show the two versions to be about the same speed, 
but in some cases 5 is faster than 3 (due to subtleties in going from 
Numeric to numpy for array support).

A good way to test is like this:

from time import clock

t0 = clock()
.....
t0 = clock()-t0
print t0

Bruce Sherwood

Gary Pajer wrote:
> I've finally upgraded to 5.0 on my XP machine.
> 
> I have a subjective sense that it might be slower than 3.0.    Before
> I try to reinstall 3.0 to check, I thought I'd ask ...
> 
> What's known about the speed of 5.0 vs 3.0 ?
> 
> -Gary
> 
> ------------------------------------------------------------------------------
> Create and Deploy Rich Internet Apps outside the browser with Adobe(R)AIR(TM)
> software. With Adobe AIR, Ajax developers can use existing skills and code to
> build responsive, highly engaging applications that combine the power of local
> resources and data with the reach of the web. Download the Adobe AIR SDK and
> Ajax docs to start building applications today-http://p.sf.net/sfu/adobe-com
> _______________________________________________
> Visualpython-users mailing list
> Vis...@li...
> https://lists.sourceforge.net/lists/listinfo/visualpython-users

Re: [Visualpython-users] visual 5.0 speed ?

[Guy K. K. <g....@ma...>]

On Tue, 10 Feb 2009 06:40:00 Gary Pajer wrote:
> I have a subjective sense that it might be slower than 3.0.

I have experienced the same thing. Particularly for huge sets of objects to be 
visualised. I'm using VPython to visualise results of numerical simulations 
and computations with potentially thousands to 100s of thousands objects.

So for some I've switched from rendering spheres to points which hugely 
improved performance again.

But I guess my use case is less usual, so it may very well be that for 
"normal" use cases 5.0 is faster than v 3.

Guy

-- 
Guy K. Kloss
Institute of Information and Mathematical Sciences
Te Kura Pūtaiao o Mōhiohio me Pāngarau
Room 2.63, Quad Block A Building
Massey University, Auckland, Albany
Private Bag 102 904, North Shore Mail Centre
voice: +64 9 414-0800 ext. 9585   fax: +64 9 441-8181
eMail: G....@ma...  http://iims.massey.ac.nz

Re: [Visualpython-users] visual 5.0 speed ?

[Bruce S. <Bru...@nc...>]

Ah. I think it is the case that the Visual 5 spheres are rendered more 
carefully than those in Visual 3 (whose spheres often looked a bit 
angular), so I can imagine that with a very large number of spheres 
there would be a penalty. This is a good example of how very specific 
applications can be either faster or slower.

Bruce Sherwood

Guy K. Kloss wrote:
> On Tue, 10 Feb 2009 06:40:00 Gary Pajer wrote:
>> I have a subjective sense that it might be slower than 3.0.
> 
> I have experienced the same thing. Particularly for huge sets of objects to be 
> visualised. I'm using VPython to visualise results of numerical simulations 
> and computations with potentially thousands to 100s of thousands objects.
> 
> So for some I've switched from rendering spheres to points which hugely 
> improved performance again.
> 
> But I guess my use case is less usual, so it may very well be that for 
> "normal" use cases 5.0 is faster than v 3.
> 
> Guy
> 

Re: [Visualpython-users] visual 5.0 speed ?

[Guy K. K. <g....@ma...>]

On Tue, 10 Feb 2009 08:02:07 Bruce Sherwood wrote:
> Ah. I think it is the case that the Visual 5 spheres are rendered more
> carefully than those in Visual 3 (whose spheres often looked a bit
> angular), so I can imagine that with a very large number of spheres
> there would be a penalty.

Possible. But I've switched first from spheres to cubes. I needed that as I 
wanted to give them individual colour to see where in colour space they'd be 
located (my application computations are 3D to 3D colour adaptation 
transformations, and that gets quite messy, so position in space were the 
first 3D coordinates, colour tuples represented the second/target 3D 
coordinates). But for my own purposes I can get away with rendering points, 
for publications then I'll have to go and code up a histogrammer that will 
aggregate voxels to 3D "summary" objects for rendering.

> This is a good example of how very specific
> applications can be either faster or slower.

Indeed. It can be *very* dependent on the specific use cases, as so often.

Guy

-- 
Guy K. Kloss
Institute of Information and Mathematical Sciences
Te Kura Pūtaiao o Mōhiohio me Pāngarau
Room 2.63, Quad Block A Building
Massey University, Auckland, Albany
Private Bag 102 904, North Shore Mail Centre
voice: +64 9 414-0800 ext. 9585   fax: +64 9 441-8181
eMail: G....@ma...  http://iims.massey.ac.nz

Re: [Visualpython-users] visual 5.0 speed ?

[Stef M. <s.m...@ru...>]

hello,

sorry for reacting so late,
but the core of my VPython wrapper was under construction.

Strange, but despite all the positive sound about speed of VPython-5,
I've the opposite experience, Vpython-5 is often 30% slower than Vpython-3,
I wonder why I seems to be the only guy experiencing that ?

I limit the fps at 50 fps.
and here are the measurements (all programs found on the web):

VP3    VP5     Program
50      30     Cherry tree
50      30     Clock
15      10     Fountain
15      10     Alfabet
50      35     Pipes
50      50     Multi-Pendulum
48      48     ODE
48      48     Double-Pendulum
50      35     Mandelbrot  Visual Part takes 12Sec/VP3   15Sec/VP5

cheers,
Stef

Gary Pajer wrote:
> I've finally upgraded to 5.0 on my XP machine.
>
> I have a subjective sense that it might be slower than 3.0.    Before
> I try to reinstall 3.0 to check, I thought I'd ask ...
>
> What's known about the speed of 5.0 vs 3.0 ?
>
> -Gary
>
> ------------------------------------------------------------------------------
> Create and Deploy Rich Internet Apps outside the browser with Adobe(R)AIR(TM)
> software. With Adobe AIR, Ajax developers can use existing skills and code to
> build responsive, highly engaging applications that combine the power of local
> resources and data with the reach of the web. Download the Adobe AIR SDK and
> Ajax docs to start building applications today-http://p.sf.net/sfu/adobe-com
> _______________________________________________
> Visualpython-users mailing list
> Vis...@li...
> https://lists.sourceforge.net/lists/listinfo/visualpython-users
>
>
>   

Re: [Visualpython-users] visual 5.0 speed ?

[Stef M. <s.m...@ru...>]

hello,

I tried the same program on another system today,
which should be faster than the system I used yesterday
(faster processor, Intel instead of AMD, faster (external) graphics card),
but it seems to be about twice as slow as my yesterday system.
On that theoretical faster but in practice slower system,
indeed the performances are about equal for VP3 and VP5 ???

cheers,
Stef

Stef Mientki wrote:
> hello,
>
> sorry for reacting so late,
> but the core of my VPython wrapper was under construction.
>
> Strange, but despite all the positive sound about speed of VPython-5,
> I've the opposite experience, Vpython-5 is often 30% slower than Vpython-3,
> I wonder why I seems to be the only guy experiencing that ?
>
> I limit the fps at 50 fps.
> and here are the measurements (all programs found on the web):
>
> VP3    VP5     Program
> 50      30     Cherry tree
> 50      30     Clock
> 15      10     Fountain
> 15      10     Alfabet
> 50      35     Pipes
> 50      50     Multi-Pendulum
> 48      48     ODE
> 48      48     Double-Pendulum
> 50      35     Mandelbrot  Visual Part takes 12Sec/VP3   15Sec/VP5
>
> cheers,
> Stef
>
> Gary Pajer wrote:
>   
>> I've finally upgraded to 5.0 on my XP machine.
>>
>> I have a subjective sense that it might be slower than 3.0.    Before
>> I try to reinstall 3.0 to check, I thought I'd ask ...
>>
>> What's known about the speed of 5.0 vs 3.0 ?
>>
>> -Gary
>>
>> ------------------------------------------------------------------------------
>> Create and Deploy Rich Internet Apps outside the browser with Adobe(R)AIR(TM)
>> software. With Adobe AIR, Ajax developers can use existing skills and code to
>> build responsive, highly engaging applications that combine the power of local
>> resources and data with the reach of the web. Download the Adobe AIR SDK and
>> Ajax docs to start building applications today-http://p.sf.net/sfu/adobe-com
>> _______________________________________________
>> Visualpython-users mailing list
>> Vis...@li...
>> https://lists.sourceforge.net/lists/listinfo/visualpython-users
>>
>>
>>   
>>     
>
> ------------------------------------------------------------------------------
> Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco, CA
> -OSBC tackles the biggest issue in open source: Open Sourcing the Enterprise
> -Strategies to boost innovation and cut costs with open source participation
> -Receive a $600 discount off the registration fee with the source code: SFAD
> http://p.sf.net/sfu/XcvMzF8H
> _______________________________________________
> Visualpython-users mailing list
> Vis...@li...
> https://lists.sourceforge.net/lists/listinfo/visualpython-users
>
>
>   

Re: [Visualpython-users] visual 5.0 speed ?

[Gary R. <gr...@bi...>]

I haven't checked this carefully but I had the impression on my system 
that 5.0 was slower too, especially with examples like drape.py but that 
playing with the value in any rate() calls affects this. The rate 
function in 3.0 used to behave as advertised. In 5.0 it seems to me that 
setting it too high actually slows down the animation loops.

Gary R.


Stef Mientki wrote:
> hello,
> 
> I tried the same program on another system today,
> which should be faster than the system I used yesterday
> (faster processor, Intel instead of AMD, faster (external) graphics card),
> but it seems to be about twice as slow as my yesterday system.
> On that theoretical faster but in practice slower system,
> indeed the performances are about equal for VP3 and VP5 ???
> 
> cheers,
> Stef

Re: [Visualpython-users] visual 5.0 speed ?

[Gary R. <gr...@bi...>]

I think Bruce meant the following to go to the list (sorry Bruce if it 
wasn't):

--

Actually, it was also the case in Visual 3 that setting the rate too
high could slow down the animation loop. Probably depends on the details
of the particular program.

It would be useful for those interested in these issues to post
relatively simple examples with compute times, and say what the
hardware/software environment is. The standard scheme looks like this:

from time import clock
....
t = clock()
.....Do Computations
t = clock()-t
print t

Make more than one run with a program and average the results because
you'll observe fluctuations, presumably associated with other random
stuff the operating system is doing. In my own measurements of this kind
I've seen little difference between Visual 3 and Visual 5.

Bruce Sherwood

Gary Ruben wrote:
> I haven't checked this carefully but I had the impression on my
> system that 5.0 was slower too, especially with examples like
> drape.py but that playing with the value in any rate() calls affects
> this. The rate function in 3.0 used to behave as advertised. In 5.0
> it seems to me that setting it too high actually slows down the
> animation loops.
> 
> Gary R.
> 
> 
> Stef Mientki wrote:
>> hello,
>> 
>> I tried the same program on another system today, which should be
>> faster than the system I used yesterday (faster processor, Intel
>> instead of AMD, faster (external) graphics card),
>> but it seems to be about twice as slow as my yesterday system. On
>> that theoretical faster but in practice slower system, indeed the
>> performances are about equal for VP3 and VP5 ???
>> 
>> cheers, Stef