LOSTCIRCUITS

3ds max 5

3ds max 5 has its own set of animated and rendering benchmarks, most of which are fairly straightforward and test one or two very specific capabilities of the graphics adapter. One interesting aspect of the benchmarks is that often, there is no screen update other than for the animated portions of the scene. This is particularly relevant in e.g. "Dual Planes" where the ball is moving between the two planes from left to right, otherwise, there would be no way for the Wildcat VP99 to reach 480 fps since that would be way beyond the fill rate limitation of that card. It is very easy to test this by opening e.g. a FRAPS window - during the next pass after closing the window, the ball will partially erase it, wherever it passes over the window. Bottom line is that some of the OpenGL applications abide by different rules than what most of us know from D3D applications.

4Viewports

All four viewports are rendered simultaneously.

Frames per second, higher is better.

8 Direct

Frames per second, higher is better.

8Omni

Frames per second, higher is better

      

Frames per second, higher is better. Using the MaxTreme drivers, this scene is not rendered correctly by the Quadro, the light effects look like a mixture of Omni and Spot lights (right picture). Therefore, even though the Quadro4 XGL wins here, we have to disqualify this result and hand it to the FireGL-X1.
Correction: the 8Spot scene in 3ds max contains four upper and four lower lights. As we were informed, the four upper lights are flagged to do a hybrid lighting effect, that is, a combination of Omni and Spot light, whereas the four lower lights are doing spot-lighting only. For some reasons, the standard OpenGL drivers ignore the "hybrid" flag, the same goes for the ATI and 3DLabs drivers. As a result, the rendering load on the Quadro 980XGL using the MAXtreme drivers is higher, which is reflected in the lower fps.

Overall, the majority of benchmarks is swept by the FireGL-X1, but this is only the first part of a whole series of benchmmark. What is almost more important here, though, is the difference that the different drivers or rather, the special "optimized" MaxTreme drivers make. Keep in mind that in some instances, the Maxtreme drivers cause a performance hit compared to the standard 43.09 drivers used here, which, however, is caused by the additional rendered effects. Keep in mind also that the MAXtreme drivers are the only ones that do the "correct" rendering.

Next Page:    => 3ds max 5 continued =>