Tech-Hounds.com

Conclusion: System and Graphics Benchmarks We will Use

Graphics benchmarks:

Call of Duty - Dawnville timedemo1

Although this game still uses the aging Quake III Team Arena engine, Infinity Ward have made their own improvements to it, making it one of the best looking FPS available. In general, you still get very high fps even with the setting turn all the way up. This game really does scale well with increases in processors' and graphics cards' performance. Just look at the resolution scaling graphs that shows a 5 fps difference for both average and minimum fps. Its also very sensitive to graphical features such as FSAA and anisotropic filtering.

F1 Career Challenge - Custom Replay

This game is probably a good example that simulation games can be used both as a system and graphics benchmark. You can see why by looking at the resolution and processor scaling graphs. While this game doesn't heavily uses shaders, it does features lots of effects and high resolution textures. This makes it ideal for testing the performance of newer graphics card. Being a racing simulation, we use it quite a lot to test anti aliasing and anisotropic filtering implementations. It doesn't have problem with FSAA like games that uses shaders (Splinter Cell, Halo, Full Spectrum Warrior).

Splinter Cell - Caspian Oil Refinery Beyond3D Demo

Since it heavily uses shaders and effects, Splinter Cell is a good graphics benchmark. The Beyond3D demo uses the Caspian Oil Refinery level because this is the one level that shaders are most heavily used. As you can see from the resolution and detail scaling graphs, they made the right decision. But Splinter Cell does have issues such as incompatibility with FSAA and we're looking to Splinter Cell Chaos Theory to replace this benchmark or at least complement it. Nevertheless, since games are now moving from traditional multitexturing and blending effects to shaders, this benchmark gives us some idea of how powerful the shader unit of graphics cards available today.

Nascar 2003 Season - Custom Replay

Continuing what we've said before, one of the reasons this game scale so well graphically is because it also features lots of effects and high resolution textures. Being a racing simulation, this game is also very good to test anti aliasing and anisotropic filtering implementations. Particularly since the Daytona track we use have lots of slanted surfaces.

A note regarding Doom III and Half-Life 2

As of this moment, we will not be using Doom III and Half-Life 2 as our graphics benchmarks. There are several reasons for this and none of them are related to different rendering methods the game applies to graphics card from different vendors. Doom III will lock the frame rate when you're playing the game to a maximum of 60 fps, so this means you will never see frame rates above that when you're playing. However, when benchmarking using timedemos, the game will render the scene as fast as possible. So the results from the benchmark doesn't really reflect or represent gameplay as far as average and maximum fps are concerned. As for Half-Life 2, the game applies same frame rate lock at 70 fps. There are ways around this, but the main reason we will not be using Half-Life 2 at the moment is its various bugs and glitches. Since Valve only distributes updates and patches through their Steam distribution network, it represent a problem for us because they can not be downloaded by any other means. So, rather than giving users and our readers an incomplete evaluation, we refrain from using these two game as benchmarks for the time being.

[Previous Page]
[Go to top]
[Next Page]