Different benchmark question

[Mloot]

Has anyone ever used the Dronez OpenGL benchmark? The benchmark gives a minimum and max. FPS and then the overall average FPS for the whole test. What I was curious about was that when I ran the test on an MX440, the min/max FPS were about 48/62, with the average being about 58fps. However, the same test on a 9000 gave min/max results of about 18/125, with the overall average being around 55fps. Is there a reason the MX card ran the test smoothly and with little variance and the 9000 had wildly fluctuating scores?

 

[Mingon]

The mx does not full support all of the directx 8 features that dronez uses. It is likely that either the cpu is doing some of the vertex shader effects or they are totally ignored. Whereas the radeon because it can do all of the effects in hardware it can both run faster - but also slower as in effect it is doing more work. Can you disable vertex shader support (i am sure it is the vertex support i could be wrong though)? this would give a more fair comparisson. Also the geforce 4mx has a fixed TnL unit whereas the Radeon's is programmable - this can allow certain games to run quicker on older cards if the TnL functions are fully supported.