I render a terrain from heightmap data. I draw little strips of terrain with GL_TRIANGLE_STRIP. Now here''s the catch: That way, all vertices (except the two edges) get transformed twice. Is there some way I can tell OGL: I have this Vertex3f and I want you to transform it to the appropriate Vertex2f. I could then use the 2d coords to tell OpenGL how to rasterize. Anybody know if this is possible? Would be great: it would about halve the amount of vertex-transformations I have to do per frame. You could even remember the transformations from last frame if the camera wasn''t moved! And then just re-transform the vertices moved since last frame.

Take control by using your own matrix code and then just pass opengl the resulting points relative to the viewer.

But I don''t think this will speed it up much, since you are already using triangle strips and a lot more graphcis cards now are doing the transformations in hardware.