Date:
Wednesday, February 12, 2003
Author:
Brent Justice
Editor:
Kyle Bennett
Google +1

3DMark03 Preview

The angry onion has been peeled back to reveal a new company name and a brand new 3Dmark as well. What does the future of synthetic benchmarking technology look like?

Article Image

Introduction:

As of last December 2002, MadOnion was renamed back to Futuremark. Futuremark has also re-designed their WebPages to a very modern and easy to navigate design. If you are not familiar with the line and benchmarks that Futuremark has created, here is a quick refresher course.

If we want to start with the very beginning we need to mention Final Reality, which was a Direct3D (DirectX 5.0) benchmark co-developed by Remedy Entertainment and VNU European Labs back in 1997. This was the very first benchmark to allude to Remedy Entertainments Max Payne which you saw the advertisement for on the side of a building in the “City Scene” 3D test. The rendering engine was an earlier engine designed by Remedy known as E2. E2 was the start that led to the Max-FX engine that later 3DMark benchmarks utilized.

3DMark99 was the first benchmark under Futuremark, which made its debut in October 1998. This benchmark was a Direct3D (DirectX 6) benchmark. It brought about many features, which reviewers were looking for to compare video cards. It introduced game tests as well as specific testing of individual features for comparison. This is the benchmark that started the whole online database of uploading your results to compare with other people online. It also introduced easy batch scripting and comparing using a result browser. A later update known as 3DMark99 MAX (requiring DirectX 6.1) was introduced in 1999, which brought about some graphics enhancements such as bump map testing. This benchmark uses the Max-FX engine as its “Real World Gaming” 3D engine. And they were right, it was a gaming engine, it was just that there were no games out using that engine except for the benchmark.

When 3DMark 2000 was introduced Futuremark changed their whole image and became known as MadOnion. Under the MadOnion name they continued their strategy to be THE Gamers Benchmark. 3DMark 2000 is a Direct3D (DirectX 7) benchmark that can take advantage of fixed function T&L capabilities in a video card. This benchmark brought with it more game tests in a low detail, medium detail and high detail settings as well as CPU tests, geometry tests, fill rate, and even memory bandwidth/AGP texture speed. The “Gaming Engine” this benchmark utilized was still the same Max-FX engine.

3DMark2001, developed by MadOnion, was the first benchmark from MadOnion to introduce tests for advanced features that were then only supported in one video card, the GeForce3, hardware vertex and pixel shaders in DirectX 8. The underlying 3D engine was again based on Remedy’s Max-FX engine. Other new features include support for DXTC/S3TC compressed textures, DOT-3 Bump Mapping, Environment Bump Mapping, FSAA. A few things that were dropped from its previous version were a CPU test, the AGP bandwidth test, and the medium quality game tests. They simply went with a low detail and high detail game test scenarios. At its graphics capability peak, 3DMark2001 is capable of rendering up to 342,000 triangles and 40 MB of 4x (160+ MB) compressed textures.

Now it is 2003, MadOnion is once again known as Futuremark and today is the day 3DMark03 is unveiled. (FutureMark site is currently getting hammered but you may try FileShack as they have multiple mirror points across the USA.)

3DMark03:

A lot has changed with 3DMark03, but is it for the better? If you noticed one thing above that stayed the same throughout each 3DMark version was the fact that they were all based on the Max-FX engine. While the Max-FX engine may not be the best 3D graphics engine in the world it was at least used in one game that was relatively successful and still has a following, Max Payne.

3DMark03 is their first benchmark that does not utilize this engine but has in fact completely moved away from any sort of known gaming engine. It is this core change that makes this version most different from all previous versions. I will quote from their Whitepaper:

“3D Engine. Previous versions of 3DMark used the MAX-FX 3D engine. The trend in 3D engines is moving towards very lightweight DirectX wrappers as more and more work is transitioned to the API and graphics card. For 3DMark03, we have based the benchmark directly on top of the DirectX 9.0 3D platform. This allows the benchmark to be independent of specific technologies embedded in a single 3D engine implementation. We have developed a set of lightweight wrapper routines on top of DirectX to aid code re-use.”

More about this in the help file as well:

“Previous 3DMarks have all used MAX-FX as 3D engine. The DirectX 8 introduction of vertex and pixel shaders made such massive middleware needless in 3D games for hardware with full support for DX8 or later. Therefore 3DMark03 is basically directly written onto DirectX. A small library of helper functions is used in order to avoid unnecessary rewriting of code for each test. All vertex processing is handled by vertex shaders and all pixel processing (except game test 1 and the fill rate tests) is done with pixel shaders. This makes 3DMark03 a very forward looking benchmark, and it might scale somewhat differently than many games at launch time, which still mainly rely on fixed function vertex and pixel processing. Then again, the mission of 3DMark is to give an insight into the next generation of 3D games and their performance requirements. Some information of the technology of the next generation of 3D games was available during the development of 3DMark03, but as usual, different types of games use different optimizations, and reflecting an average of all that in advance is next to impossible. 3DMark03 therefore measures above all DirectX vertex and pixel shader performance, since this is our best estimate of next generation 3D game technology and performance requirements.”

3DMark03, like its predecessors, is completely Direct3D based with no hints of any OpenGL testing. Futuremark does acknowledge that users have requested OpenGL testing.

"As Direct3D® has higher usage in games and more uniform driver support, we only support Direct3D. However, we actively monitor the progress of OpenGL adoption."