[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

Re: [pygame] [PyGame + PyOpenGL] Setting Target Surface



On 4/6/06, Peter Shinners <pete@xxxxxxxxxxxx> wrote:
> On Thu, 2006-04-06 at 21:34 -0700, Brian Fisher wrote:
> > I can't see why there would be a speed difference in terms of just
> > copying pixels, but who knows... also, there is only one accumulation
> > buffer, while you can have a lot of textures
>
> glReadPixels will stall your hardware graphics pipeline. It will make a
> noticeable speed drop.
>
Sorry, I wasn't explicit - I was referring to the difference between
using the accumulation buffer vs. using textures (between glAccum and
glCopyTexImage2D). Because they are both copying pixels between video
card memory I assume they would be similar in performance.

yes, glReadPixels (and all system memory <-> card memory ops, like
glDrawPixels, glTexImage, and glGetTexImage) are performance dogs not
suitable for real-time rendering