Benchmarks - 3d Mark 2001SE & Unreal 2003
3D Mark 2001 SESo let's see how bandwidth related to game performance, after all this site is targeted at you, the gamer. And although the difference is not something to write home about it does influence gaming somewhat. The results below where done with a GeForce4 Ti 4600. The DDR frequency looks a bit odd, but we tested it this way on purpose as changing the FSB to other values to get an exact 266/333/400 frequency would influence the benchmark results.
The Second Edition is a major enhancement to its 3DMark® benchmark solution, originally launched in 1998. Combining DirectX 8.1 support with the latest in 3D graphics, 3DMark2001 SE objectively measures how effectively a PC runs 3D graphics applications and provides users with an online results comparison tool that allows them to make informed hardware assessments and upgrade decisions
Using Remedy Entertainment's acclaimed MAX-FX Technology, 3DMark2001 demonstrates how to maximize your 3D gaming performance by using real-world gaming technology to test your system's true game performance abilities.
| 3D Mark 2001 SE | C2.5:3:3 |
| 414 MHz | 13439 |
| 387 MHz | 13231 |
| 310 MHz | 12681 |
As you can notice yourself, the difference in MHz is good for a significant change in performance on this system. Remember one important aspect here. The Processor frequency never changed at all, the diffenences and changes in the benchmark scores are 100% based on memory frequency and their changes.
Unreal 2003
The Unreal Tournament 2003 benchmark is much more modern than Quake III. A nice little addition to UT2003 is that it has a built-in benchmark utility. You c
an find the benchmark in the system subdirectory for the game after it installs. The file is appropriately named 'benchmark.exe'. Run this file, select the resolution you want to benchmark, and it will go through 4 different demos, 2 are "flyby" and 2 are "botmatch" demos. The flyby demos give you an idea of how high frame rates would be without bots, while the botmatch demos give you an idea of what kind of fps you can expect while actually playing the game. The benchmark actually leverages more cpu and graphics intensive technologies that the engine is capable of.
In this Unreal2 benchmark we'll take a somewhat different approach, we again raise the frontside bus of the PC towards 155 MHz and are going to check the difference between RAM timings to proof to you that a game indeed is sensitive to bandwidth speeds and timings also.
| Unreal 2003 | FlyBy | Bothmatch |
| 414 MHz 2 : 7 : 2 : 2 | 166.8 | 62.8 |
| 414 MHz 2.5 : 7 : 3 : 3 | 164.8 | 62.2 |
| 310 MHz 2 : 7 : 2 : 2 | 162.3 | 61 |
Again we can make note of the fact that differences in memory timings and frequency have an impact on gaming performance. Okay agreed, it's not a huge difference in particular with Unreal 2003 but then again every little bit helps if you want to build that ultimate performaing gaming-rig right ?
