OpenGL Geometry Benchmark was designed to benchmark and test the geometry Transform & Lighting power of different video cards under OpenGL. Project has been abandoned.

The program can test many combinations of geometry formats, and also use the NVidia and ATI extensions that provide access to video or AGP memory. This is later referred as a static vertex array ( Video ) or a dynamic vertex array ( AGP ). It's possible to use interleaved or flat ( = streaming ) arrays, no or 3 lights to do a light-stress, the kind of primitives ( lists as opposed to strips ), but also the type of indices ( short indices vs. long indices ), and the tesselation of the scene ( geometry-stress). Last but not least, you can enable/disable the methods you want to test, which can be useful if you just want to test a specific combination. Please note that, by default, all the tests will be performed. Each test lasts 3 seconds exactly; the full test will run for 10 to 15 mins.

After the tests are done, you'll be asked if you want to send a log file to the website. I'm planning to compile a list of results on that page, so if you want to help, just click Ok :) If you prefer to cancel, nothing will be sent. Finally, a text window, using your default text editor will open, with the results. You'll find some useful information ( CPU type, CPU speed, OpenGL renderer, and extensions ), then all the results, sorted by MTS ( millions of triangles per second ). Other notations include FPS ( frames per second ). The flags, after the test name, are as follow s:

High tesselation vs. Low tesselation ; 0 lights vs. 3 lights; Interleaved format vs. Flat format; Strips vs. Lists; Short indices vs. Integer indices.