Suppose I modeled two cubes. The first one has all sides looking like this, just a flat simple square polygon. Then I made another cube identical to the first one, but this time the sides look more like this, a square polygon with a diagonal edge.

My question is, which of the two boxes would be cheaper for SWBF2 to render? Which of the two approaches is less system intensive to show ingame? And why?

I could be wrong, but I think I remember reading a long time ago that VM triangulates your model when you munge it, so if that's true then the first shape you showed would just end up looking like the second one before it's even handled by the game engine.

There's no way for the graphic card to easily understand anything but triangles. So the short answer is that triangles are more easy to render, because truly rendering a quad would require you to create your own renderer.

The way graphic cards process polygonal meshes, the format pretty much all games use, is (and I am simplifying here) by receiving them as a large strip of triangle definitions. I could use more nuance, but essentially it's like that because the math formulas to describe forms in 3d space are defined this way. sortof
Here's a video where a more qualified person describes what's going in, if you're interested:
http://youtu.be/KdyvizaygyY

VM triangulates the models automatically. However, it's really bad at it. Smaller quads usually just disapear when VM converts them.

tl;dr always triangulate your models before exporting it to msh. Working in quads is the standard, just don't export that way.

Thanks a lot for all the info guys.

And just to clarify everything in my head, a model should be triangulated before exporting to prevent Visual Munge from doing it in the not best way, right?

And great video mswf. Thank you very much for the link.

Correct.

Allright. Thanks a lot.