Comments on: GPU Geometry Map Rendering – Part 2

By: Tristam MacDonald

Tristam MacDonald — Sat, 15 Oct 2011 16:44:45 +0000

So you have probably long ago figured out the reason why texture coordinates don’t interpolate to 1.0, but in case anyone else stumbles upon it:

The issue is that the pixels in your output texture have non-zero size, and they are specified by a particular point (the top-left of the pixel, if your coordinate system is standard).

Texture coordinates start at 0.0 on the left hand side, and since pixels are specified by their left-hand side as well, the first pixel falls at 0.0 as expected.

On the right-hand side of the texture, the texture coordinate is 1.0. But since pixels are specified by their left-hand side, the pixel at 1.0 is actually one past the end of the texture. Thus the last valid pixel is actually at coordinate (1.0 – pixel width).

By: dallas

dallas — Fri, 03 Dec 2010 14:06:30 +0000

Great post! Nice!

By: Dave

Dave — Sat, 24 Jul 2010 20:05:28 +0000

I’ve never liked OpenGL much, and besides, I wouldn’t be able to create planets on the Xbox . C# and XNA are a very powerful combination, and suit my needs very nicely.

By: Jake

Jake — Fri, 23 Jul 2010 23:50:16 +0000

Have you tried using OpenGL, or something which is probably more suitable, like OpenCL?

This would give cross-platform compatibility, and might work a little more predictably.

By: GPU Geometry Map Rendering, Continued « Sgt. Conker

GPU Geometry Map Rendering, Continued « Sgt. Conker — Tue, 12 Jan 2010 11:27:13 +0000

[...] Dave Carlile strikes again with a continuation of his first post about GPU Geometry Map Rendering. This time, he goes through the whole process of debugging his shader using PIX, and finding a solution to his problems. Read the whole post on his site. [...]