Test results and analyses, summing-up

Scene 4

The forth and the fifth scenes show performance of the accelerators in operation with light sources:

MAXTreme:

Well, the gaming cards are superb. The full victory and the MAXTreme driver do not help the the professionals to catch up with them.

The same situation takes place in the fifth scene:

MAXTreme:

The last test shows whether the 3D MAX and the accelerators work correctly when the resolution is changed.

• Radeon 8500 didn't suffered from changing the resolution from 1280x1024 to 1600x1200 at 32-bit color and back.
• FireGL2 wasn't able to restore the image when the resolution changed from 1280x1024 to 1600x1200 and back at 32-bit color:
  

• Quadro 2 MXR wasn't able to withstand changing of the resolution under the OpenGL driver, however with the MAXTreme it was Ok:


• Quadro 2 PRO wasn't able to withstand changing of the resolution under the OpenGL driver either, however with the MAXTreme it was Ok:

• Quadro DCC suffered from changing the resolution under the standard OpenGL driver on the Pentium4 platform, however, the image didn't changed under the MAXTreme:

• GeForce4 performed excellently when the resolution changed from 1280x1024 to 1600x1200 and back at 32-bit color.

And now you can look at the screenshots of our participants. As there is no big difference between the quality of the screenshots of these and the previous tests we are showing just some of them:

GeForce4:

Quadro 2 PRO:

Quadro DCC:

Quadro 2 MXR:

Conclusion

I think that the most optimal solution among professional cards for the most convenient work will be the Quadro DCC, however its price is quite high, especially considering that the GeForce4 shows the same or even higher results. The Quadro 2 Pro is a good choice as well. The worst purchase for professional 3D modeling is ATI's Radeon 8500; and although the Quadro 2 has worse results than the Radeon 8500, it has less errors. Probably, it's because of the drivers. But we should be loyal as the card belongs to a different market sector. The FireGL2 performs well in these tests.

As far as the platforms are concerned, I can't say that the one is certainly better than the other. However, we should allow for megahertz. The Pentium4 is stronger in this parameter: 2100 MHz against 1533 MHz. However, the results are very close in the tests. But the Athlon based system is much cheaper and has some more reserve.

Soon we are going to carry out tests on the 3D MAX ver.4 which is optimized for the SSE. We'll see whether the Athlon will be able to win the battle...

Write a comment below. No registration needed!