Sign in to follow this  
gunning

Wanted: Performance Tips for Reading Back Render Target Data

Recommended Posts

A graphics algorithm I am using requires that I render scene depth information to a render target and then read it back to the CPU. That's all great, but it drops my framerate dramatically. In fact, my framerate goes from 2000+ to around 100. This is understandable since the CPU must flush the graphics commands, the GPU must respond to it (possibly having to render the scene), and then the CPU must wait for the texture data. My question is: What tricks can I use to speed this up? I've currently tried reducing the dimensions of the render target by half (currently 1024x768), and that boosts the framerate to about 300, but I haven't determined if my graphics algorithm will still work with a reduced render target and it's not an optimization I'm able to make right now. So ... experts ... any advice?

Share this post


Link to post
Share on other sites
Can't you do this through a shader? Without really leavning the shader, avoiding roundtripping between CPU and GPU or DirectX and Shaders...

Share this post


Link to post
Share on other sites
Quote:
Original post by leidegren
Can't you do this through a shader? Without really leavning the shader, avoiding roundtripping between CPU and GPU or DirectX and Shaders...
Unfortunately, no. The algorithm I am using requires depth values to generate new geometry.

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

Sign in to follow this