iXBT Labs :: Computer Hardware In Detail

Platform

Video

Multimedia

Mobile

Other

Professional 3D Accelerators in the 3D Studio MAX 4.26: March-April 2002

Performance results. Quantitative tests. Anti-Aliasing. Addition.

About addition

In this section I will estimate operation of the MAXTreme driver. The driver's version with which I carried out the tests before was too imperfect, and I got a new version with the errors corrected. The test scene remains the same, as well as the test conditions; the only thing that changed is the driver's version. Well, let's see what the drivers can do.

I got some explanations from ELSA concerning a too low performance of the whole Quadro series in the anti-aliasing mode with the MAXTreme driver used (you can read them in the 4th part of the review); but it turned out that the problem is not simple to solve. The Use Triangle Strip option has nothing to do with it, because the rendering is processed not by means of the video card but by the processor's means, whether this option is enabled or not. When this option is disabled, the rendering becomes faster because half of the lines disappear, but it doesn't influence quality or quantity. However, we had all this in the last version, and today we have a new one - 4.00.21. Let's see whether the speed is improved:

Old version - 4.00.13

New version - 4.00.21

Well, the programmers have managed to improve the speed. Now the Quadro4 900XGL is the leader, as it must be, then we have Quadro4 750XGL, which is followed by the Quadro2 PRO, Quadro DCC and the last one is Quadro2 MXR. Let me explain the layout - look at the frequencies of the chip and of the memory.
Chip:

  • Quadro4 900XGL: 300 MHz
  • Quadro4 750XGL: 250 MHz
  • Quadro2 PRO: 250 MHz
  • Quadro DCC: 200 MHz
  • Quadro2 MXR: 200 MHz

Memory:

  • Quadro4 900XGL: 650 MHz
  • Quadro4 750XGL: 550 MHz
  • Quadro2 PRO: 400 MHz
  • Quadro DCC: 460 MHz
  • Quadro2 MXR: 183 MHz

Now, I think, you understand why we have such results. In this scene the most part of the load is laid on the chip. The faster the chip and memory of the card, the faster the rendering. This is just a rough approximation because I don't take into account different cores, different throughput of buses, different memory and other factors. But the main thing is that this scene loads nothing but the chip and the memory, and the task of the card is very simple. And now let's enable anti-aliasing.

Here are the respective settings in the drivers (for the Quadro line):

Well, the AA is enabled, let's look at the results in the new version.

Old version - 4.00.13

New version - 4.00.21

No we can see that the AA is processed by the GPU, while in the previous version it was done by the CPU. The error is corrected. Note that the speed decrease is about 2% on the fastest card - Quadro4 900XGL. This is not much. The driver works fast. But the speed is not enough for operation with the line anti-aliasing enabled. AA quality is also important.

I'm not going to show screenshots of the two versions without the AA, because they do not differ in rendering quality. I will try to compare these versions with the AA enabled.

Quadro2 MXR

First comes the screenshot with AA in the ver.4.00.13, then with AA in the ver.4.00.21.

Here are enlarged pictures:

Line antialiasing enabled, 4.00.13


Line anti-aliasing enabled, 4.00.21

Well, the AA is of rather high quality now. It concerns all Quadro cards:

Quadro2 PRO

First comes a screenshot with the anti-aliasing in the 4.00.13, then with the AA in the 4.00.21.

Line anti-aliasing enabled, 4.00.13


Line anti-aliasing enabled, 4.00.21

Quadro DCC

First comes a screenshot with the anti-aliasing in the 4.00.13, then with the AA in the 4.00.21.

Line anti-aliasing enabled, 4.00.13


Line anti-aliasing enabled, 4.00.21

Quadro4 750XGL and Quadro4 900XGL

I joined these two cards because they differ only in the speed of operation, and they are identical in quality. First comes a screenshot with the anti-aliasing in the 4.00.13, then with the AA in the 4.00.21.

Line anti-aliasing enabled, 4.00.13


Line anti-aliasing enabled, 4.00.21

Well, the new version of the MAXTreme driver has improved both speed results and quality. The anti-aliasing now works on the video card level, and it works well. The drawback is that anti-aliasing is still rough, as we can see on the enlarged fragments: the lines are rather think and the stairs are still noticeable sometimes. If the programmers managed to achieve the results close to those of the 3DLabs' card, the Quadro cards will be unequalled both in speed and in quality of rendering.

Article navigation:



  Most Popular Reviews More   RSS  

Intel Core i5-660/661 Processors

The long-awaited dual-core solutions for the LGA1156 platform.
Jan 09, 2010 · Processors: Intel

AMD Athlon II X3 425, 435 Processors

Three cores for the price of two.
Oct 30, 2009 · Processors: AMD

E-MU 0404 USB

A top-class external audio interface.
Nov 23, 2006 · ProAudio

Intel H55, H57 Chipsets

Integrated chipsets for lower-end Nehalem processors.
Jan 18, 2010 · Chipsets

Creative Sound Blaster X-Fi Surround 5.1

An external X-Fi solution in tests.
Sep 09, 2008 · Sound Cards
  Latest Reviews More   RSS  

ASUS M4A785G HTPC/RC Motherboard

A microATX board with a high-quality audio codec and a remote.
Feb 05, 2010 · Motherboards

Thermaltake Element G Mid Tower Case

Compact size, fan controller, and stylish lights.
Feb 03, 2010 · General Platform

ATI RADEON HD 5670 Graphics Card

A low-end solution with DirectX 11 support.
Feb 02, 2010 · Video cards: ATI GPUs

i3DSpeed, January 2010

Updated the list of benchmarks. Added test results for Radeon HD 5670.
Feb 01, 2010 · 3Digests

Intel Celeron E3300 Processor

A full-fledged low-end solution or a "socket stub?"
Jan 29, 2010 · Processors: Intel
  Latest News More   RSS  
 
  Useful Links Get listed  

Platform  ·  Video  ·  Multimedia  ·  Mobile  ·  Other  ·  Compare Prices  ||  Feedback  ·  Advertise at iXBT Labs  ·  About us  ·  Affiliates  ·  Forum


52

Copyright © Byrds Research & Publishing, Ltd., 1997—2010. All rights reserved.