To answer the first question we need to look at the performance results using real world application based benchmarks. I’m going to assume the system is used for a wide range of applications, from games to business applications, with a mix of new and some not so new… but for the most part they are up to date.
I’ve chosen the latest Winstone series from eTesting Labs(Ziff Davis), Business Winstone 2001 and Content Creation 2002, because I feel they do good job of providing a sampling of commonly used and up to date programs running just as they would for the user. Business Winstone achieves its score by testing the following programs:
- Five Microsoft Office 2000 applications (Access, Excel, FrontPage, PowerPoint, and Word)
- Microsoft Project 98
- Lotus Notes R5
- NicoMak WinZip
- Norton AntiVirus
- Netscape Communicator
Content Creation Winstone uses a different mix:
- Adobe Photoshop 6.0.1
- Adobe Premiere 6.0
- Macromedia Director 8.5
- Macromedia Dreamweaver UltraDev 4
- Microsoft Windows Media Encoder 7.01.00.3055
- Netscape Navigator 6/6.01
- Sonic Foundry Sound Forge 5.0c (build 184)
Between those two benchmarks we should be able to get a pretty good idea of the performance difference to expect under different kinds of system usage, from the traditional office applications to the newer multimedia applications.
Along with the two Winstone benchmarks I’d like to run some benchmarks that will help to identify the effect that changing the CPU, memory and Mainboard will have on the different subsystems. In order to do that I’ll use the Winbench 99 series of tests. Unlike eTesting LabsI feel it’s important to still use the CPU and FPU scores, so I’m using an older version that still includes them. The following subsystems will be tested:
- CPU (includes not only the CPU but also any CPU cache & main memory performance)
- Disk (both business and high-end applications)
- Graphics (both business and high-end applications)
To round out the tests I’ll use the following applications:
- 3D Mark 2001 – a 3D gaming benchmark that includes results from a number of different test at different resolutions.
- Video 2000 – a benchmark for video clip playback.
- Quake – a game in which its FPS scores reflect both CPU power and memory performance.
- Unreal Tournament – a more up to date game with a bit different system usage.
But instead of looking at confusing test score numbers I am going to first look at what really counts – the performance gain in percentage after I upgrade the CPU, Mainboard and memory. It may sound impressive to hear that one system is 20 FPS than another, but if it was stated as being 5% faster that wouldn’t be so impressive. In some cases a score that is 200 points higher in a benchmark could be less than 5% (such as 3D Mark 2001)! As a rule of thumb, what I look for is a difference of at least 10% to even be noticeable to the user. Let’s see if the iP4 CPU with an i845 chipset based Mainboard using DDR memory will be sufficiently faster than an iP3 and i815 chipset with SDRAM to even notice.
One thing I did not do was use any non-standard setting for the tests, such as drivers or utilities that would raise the scores above what the user would see in normal use. So the test results in some cases may be limited by other hardware and not show the true potential of the iP4 CPU (that may be the subject of an article another day), but they will show how it will perform in normal use.
Be the first to discuss this article!