This review starts a new subsection in testing of professional video cards produced for a sector of professional 3D graphics. We will engage video cards from a gaming sector as well because a lot of studios use gaming cards for 3D modeling. Although their range of application is not wide, we won't disregard this aspect.

One of the most popular packets in the professional 3D graphics sphere is 3D Studio MAX from Kinetix (subdivision of Autodesk). We will start our tests with this packet. Further, with a more experience and improved testing techniques, we will include more powerful and advanced programs.

First of all, we will analyze operation of 5 accelerators in the 3D Studio MAX: 3 professional cards and 2 ones from a gaming sector.

Cards

NVIDIA Quadro2 Pro

The Reference card NVIDIA Quadro2 Pro card has an AGP x2/x4 interface, 64 MBytes DDR SDRAM located in 8 chips on the right side of the PCB.

Brief technical characteristics:

• Graphics Core: 256-bit
• Triangles per Second: 31 Million
• Pixels Per Second: 1.0 Billion
• Memory Bandwidth: 6.4 GBytes/s
• NVIDIA Shading Rasterizer (NSR)
• Second Generation Transform and Lighting (T&L)
• Unified Driver Architecture (UDA)
• 32-bit Z/Stencil Buffer
• 64 MBytes Unified Frame Buffer
• AGP 4X with Fast Writes

The NVIDIA Quadro2 Pro serves a base for only one card - ELSA Gloria III which is a reference card from ELSA; it comes with special software, including a driver for operation with some packets for 3D modeling - MAXTreme (it is freely available at the NVIDIA's site).

NVIDIA Quadro2 MXR

The Reference card NVIDIA Quadro2 MXR card has an AGP x2/x4 interface, 32 MBytes SDR SDRAM located in 4 chips on the right side of the PCB.

Brief technical characteristics:

• Graphics Core: 256-bit
• Triangles per Second: 25 Million
• Pixels Per Second: 400 Million
• Memory Bandwidth: 2.93 GBytes/s
• NVIDIA Shading Rasterizer (NSR)
• Second Generation Transform and Lighting (T&L)
• Unified Driver Architecture (UDA)
• 32-bit Z/Stencil Buffer
• 32 MBytes Unified Frame Buffer
• AGP 4X with Fast Writes
• Digital Vibrance Control (DVC)

The NVIDIA Quadro2 MXR serves a base also for only one card - ELSA Sinergy III which is a reference card from ELSA; it comes with special software, including a driver for operation with some packets for 3D modeling - MAXTreme.

ATI FireGL2

The ATI FireGL2 card has an AGP x2/x4 interface, 64 MBytes DDR SDRAM located in 8 chips on the right and back sides of the PCB.

Brief technical characteristics:

• Graphics controller - IBM Chipset: RC 1000 256-bit Graphics Rasterizer and GT1000 hardware geometry engine with integrated features including;
• VGA Controller - 2D/3D Raster Engine and dual Texture Unit;
• 8-bit Double Buffered Overlays;
• Video Overlay Unit;
• 2 DMA / BLIT Units;
• Polygon Setup Processor;
• 300 MHz / 30-bit Palette DAC, including four color lookup tables and gamma correction table;
• 256-bit Rasterizer & DDR memory interface;
• Bus type - AGP 2X/4X v2.0 Compliant;
• Memory configuration - 64 MBytes DDR SGRAM, Unified Framebuffer;
• 3D performance:
• 27 million Triangles/second, G-Shaded, Z-buffered, non-Textured
• 31 Million Anti-Aliased Vectors/second
• 410 Million Pixels/second fill rate, G-Shaded, Z-buffered, non-Textured
• 200 Million Pixels/second Trilinear Texture fill rate (Mip-mapped)

This card was released by Diamond Multimedia yet before it amalgamated with S3 (today's Sonic Blue). When Sonic Blue decided to leave a video card sector its gaming and professional video cards were were sold to other companies. ATI Technologies became an owner of the FireGL family. Until now these accelerators were based on the IBM processors. And only starting from the RADEON 8500 the Canadian company announced a new line FireGL 8700/8800 based on its own GPU.

That is why this card has a BIOS from S3 Corporation and a software set from Diamond Multimedia.

NVIDIA GeForce3

The Gainward CARDEXpert GeForce3 PowerPack based on the NVIDIA GeForce3 has an AGP x2/x4 interface, 64 MBytes DDR SDRAM located in 8 chips on the right side of the PCB.

Brief technical characteristics:

• Graphics Core: 256-bit
• Triangles per Second: 76 Million
• Pixels Per Second: 800 Million
• Memory Bandwidth: 7.3 GBytes/s
• NVIDIA nfiniteFX Engine
• 3D Textures
• Vertex Shaders
• Pixel Shaders
• Shadow Buffers
• High-Definition Video Processor (HDVP)
• High-Resolution Antialiasing (HRAA)
• Lightspeed Memory Architecture
• Unified Driver Architecture (UDA)
• 32-bit Z/Stencil Buffer
• 64 MBytes Unified Frame Buffer
• AGP 4X with Fast Writes

There are quite a lot of experts in professional 3D graphics who use accelerators of the gaming class based on the NVIDIA processors. They work excellently in packets of the 3D Studio MAX level thanks to a debugged and multifunctional OpenGL driver.

ATI RADEON 64 MBytes DDR (183 MHz)

The ATI RADEON 64 MBytes DDR VIVO on the ATI RADEON Graphics has an AGP x2/x4 interface, 64 MBytes DDR SDRAM located in 8 chips on the right and back sides of the PCB.

Brief technical characteristics:

• Graphics Core: 256-bit
• Triangles per Second: 30 Million
• Pixels Per Second: 366 Million
• Memory Bandwidth: 5.8 GBytes/s
• CHARISMA ENGINE
• 3D Textures
• PIXEL TAPESTRY architecture
• Twin Cache Architecture
• Superscalar Rendering
• Environment bump mapping
• 32-bit Z/Stencil Buffer
• 64 MBytes Unified Frame Buffer
• AGP 4X with Fast Writes

The RADEON is a typical gaming accelerator which is not meant for professional 3D graphics and its software is not optimized for these purposes, but it was interesting for us to look at this card in the 3D Studio MAX, especially considering that there is already one gaming card on the NVIDIA and it is desirable to take a look at its competitor. Unfortunately, when we started these tests the ATI RADEON 8500 wasn't yet available (as well as the RADEON 7500), and we had to take the most powerful accelerator at that moment - RADEON 64 MBytes DDR.

Terms & conditions

Testbeds:

• Pentium4 based computer (Socket 423):
• Intel Pentium4 1500;
• ASUS P4T (i850) mainboard;
• 256 MBytes RDRAM PC800;
• Quantum FB AS HDD, 20 GBytes;
• Athlon based computer:
• AMD Athlon 1200;
• Chaintech 7KJD (AMD760) mainboard;
• 256 MBytes DDR SDRAM PC2100;
• IBM DTLA HDD, 45 GBytes;

The tests were held on two operating systems: Windows NT 4.0 Service Pack 6 and Windows 2000 PRO Service Pack 2. For a new accelerator we reinstalled the system to avoid possible errors.

The tests were carried out in two resolutions: 1280X1024 and 1600X1200 at 32-bit color. We tested the card at the lower resolution, and checked whether the 3D MAX would work in the higher one. To estimate quality of drivers and operability of the accelerators under different APIs the tests were carried out in all available APIs - OpenGL, Direct3D and APIs from the chip manufacturers.

All scenes use omnidirectional light sources.

There are 5 scenes:

1. Scene 1 consists of 3 light sources and 50 objects combined in one. The number of vertices is 37,256 and the number of triangles is 67,139. Here is a rendered screenshot of the scene:

2. Scene 2 is more complicated, it consists of 1 light source and 118 objects. The number of vertices is 306,468, and the number of triangles is 26,051. Here is a rendered screenshot of the scene:

3. Scene 3 is the most complicated; there are 564,196 vertices and 1,098,352 triangles. There are also 13 objects and one light source. Here is a rendered screenshot of the scene:

4. Scene 4 doesn't contain much geometry; it is used to test operation of accelerators with a great number of light sources. There are 4 irregular-shaped geometric objects, 24,960 triangles, 12,488 vertices and 10 light sources of the Omni types. Here is a rendered screenshot of the scene:

5. Scene 5 is identical to the previous one, except light sources which are now of the Spot type. This light type is more complicated for calculation. Here is a rendered screenshot of the scene:

In these tests we focused on estimation of the speed of operation in different variants of shading of the preliminary rendering both in a large window and in small ones. We also fixed errors which occurred during the operation. The tests used standard settings of the 3D MAX and specially developed scenes which were different in complexity.

Each scene was viewed both in a full window and in 4 modes:

• Wireframe

• Smooth + HighLights

• Smooth + HighLights, texture on, correction off

• Smooth + HighLights, texture on, correction on

and in the 4-window mode, each window having its own mode - clockwise starting from "1 o'clock" - WireFrame, Facets, Smooth+HighLights, Smooth+HighLights, texture on, correction on.

Settings for each scene can be changed in the contextual menu. All viewport parameters are set in the Viewport Configuration menu.

The visualization modes we took for the tests - wireframe, facets, smooth+highlights and smooth+highlights with texture correction - are used most of all in an everyday work:

• Wireframe - a mode of displaying geometry in the form of a wireframe. It is the most convenient for creation and editing of geometry. Although it doesn't always give an idea about a rendered scene, this mode is used most of all. If you doesn't use preliminary rendering modes at all, you should pay attention exactly to these results.
• The Facets mode displays geometry in the form of triangles of the material's color and gives an idea how figures will look like and how they will be illuminated. It is rough, but you can get it very fast.
• The most informative, or close to the final form of a scene, is the smooth+highlights mode and its variant - smooth+highlights with texture correction. Anti-aliased imaging of illuminated surfaces with textures allows us to estimate an appearance of a scene, though an illumination is rough. The mode with texture correction was used to estimate a performance drop when it gets activated in the same visualization mode.

[ Part 2 ]

[ Part 3 ]

[ Part 4 ]

[ Part 5 ]