Comments on: Triangle Optimizations https:/2008/11/triangle-optimizations/ Developer resources for the X-Plane flight simulator Tue, 01 Feb 2011 18:54:17 +0000 hourly 1 https://wordpress.org/?v=6.6.1 By: Benjamin Supnik https:/2008/11/triangle-optimizations/#comment-899 Thu, 27 Nov 2008 08:12:00 +0000 http://www.x-plane.com/dev_blog/?p=281#comment-899 Hi Rob,

The limit is 255 vertices per primitive, so you could have 85 tris per primitive. This only matters for DSF size though.

But if you do have easy access for strips, it’s not bad to use them; while the strips themselves don’t matter (they are turned to indexed tris) as of now, the only way to control vertex order is by using strips/fans. (The vertices are ordered by the strip in this case.)

So my suspicion is that your cache coherency will be better with strips than individual tris (which get sorted by, well, lord knows what).

]]> By: Anonymous https:/2008/11/triangle-optimizations/#comment-900 Thu, 27 Nov 2008 04:44:00 +0000 http://www.x-plane.com/dev_blog/?p=281#comment-900 Hi Ben, What a coincidence! I was just about to drop you an email with question about triangles in base mesh.
My tool G2xpl after converting original base mesh outputs only triangles (skips fans and strips). Each primitive (between BEGIN_PRIMITIVE and END_PRIMITIVE) contains at most 3 triangles (someone mentioned that there is some limit of number of triangles in 1 primitive, not sure if this is correct and what’s the meaning of primitive is).
You just blogged that strips don’t matter so much on modern desktop machines… Are there any other means to optimize my output mesh for better performance or I should leave it all to be optimized by DSF loader?
Thanks/Rob

]]>