1) I am trying to use the GPU hardware rendering capabilities to speed up the rendering. How do you know what features (i.e. framebuffer objects, vertex buffer objects, etc) are actually using the GPU?
2) I implemented the FBO using a texture attached to the FBO as GL_RGBA8 and a renderbuffer used as the DEPTH attachment. When I did this I saw a much higher runtime as opposed to when I rendered to a XPixmap as the current context. Any ideas why that may be?
3) The code I’m working with creates a texture and does a glCopyTexSubImage2D(). I believe that this copies from the framebuffer to the texture? Does any further rendering after that point go to the texture or just the framebuffer? It appears that the depth effect works fine if I uncomment out the section of code that allocates the texture and then does the glCopyTexSubImage2D() when used with the FBO but no depth effect is seen when I comment out the texture portion and use the FBO as well.
I was expecting to see the FBO be really fast and see the depth effect when a texture was attached to the FBO as GL_RGBA8 and a renderbuffer attached to the FBO as the DEPTH attachment.
Sent: Tuesday, June 05, 2012 9:49 AM
To: Steven Herrick
Subject: Re: Offscreen rendering with OpenGL | AMD Developer Forums