On Wed, Jan 27, 2010 at 2:44 PM, Florian Bösch <[email protected]> wrote:

> On Jan 27, 2:50 pm, Tristam MacDonald <[email protected]> wrote:
> > As I have mentioned many, many times now, it isn't OpenGL which is broken
> > here - it is wave.py.
> No mention of RGBA32F made so far.
>

I didn't detect this particular performance issue up until that point,
because I didn't want to take the time to debug the program.


> > In particular, GPU-to-GPU glReadPixels is only supported for some
> particular
> > pixel formats, and GL_RGB (which you are using) is not one of them.
> > Switching the texture formats to GL_RGBA32F, the glReadPixels mode to
> > GL_RGBA, and switching the vertex buffers to 4 elements per vertex brings
> > the application above 30 fps for a map size of 1024x1024.
> Thanks a bunch, it's true, I switched over and it copies much faster.
>
> Now if you had mentioned the GL_RGBA issue from the start, we could
> have saved ourself a lot of bytes.
>

My assertion was never that wave.py could be perform better - though it
obviously can. My assertion is that you cannot make a general statement
about the performance of OpenGL on the basis of a single test program.

It was obvious from the start that an existing game like Quake 4, or indeed
any of NVidia's OpenGL examples, manages to push more than 75 MB/s bandwidth
on your card. This immediately indicates that your program is broken - thus
my incredulity at your repeated assertion that the API itself (which has
been vetted by thousands of programmers) is broken.

-- 
Tristam MacDonald
http://swiftcoder.wordpress.com/

-- 
You received this message because you are subscribed to the Google Groups 
"pyglet-users" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/pyglet-users?hl=en.

Reply via email to