How We Test System Performance,
The Benchmark Software We Use,
And Why We Do It This Way!
How we test system performance:
In order for you to understand how we test the performance of the computers we build, you must first understand our design concepts and the philosophy behind them. The major computer manufacturers design and build their computers around an conceptual idea of a new product to entice a buyer. Each of the components used in those computers are molded around that concept. The mass manufacturer needs to be able to do this in order to mass produce large numbers of computers at the lowest possible cost in order to derive a profit.
Our design concepts are entirely different than the major manufacturers. Aside from the specials we build from time to time, every system that we design and build is built around you, the user. We start with what you want to do with your computer and what you want it to do for you. There are no preconceived ideas, no base units, no standard cases and no basic motherboards, everything begins with you. From there we choose components and mate them based upon test data that we have developed that indicates which components will perform well together. Of course, if you have preferences for certain hardware components, we will then advise you of what does and does not work well together.
Since 99.9% of our systems are custom built, for testing purposes we have prepared several base systems upon which to calibrate and standardize our testing. You can view the composition of our calibration systems by clicking here. Our test data concentrates on computers assembled from readily available brand name components that are combined to provide you with a computer that fits your needs. With this in mind, we test your machine with the benchmark software noted below and compare it with both the base calibration system as well as our advanced systems. We we do not destroy any of the test data that we develop, we use it every day to measure new technology. This makes it easier to compare systems where the only change was a video card, or other similar device.
We make every effort not to show any favoritism between component manufacturers, however in those cases where we do, it is for no other reason than the manufacturer has demonstrated better than average technical support of their product lines, as well as an admirable warranty replacement policy.
Unlike much of the test results that you will find published elsewhere on the Internet, our data is based upon everyday scenarios of how people actually use their personal computers, work stations and servers. We load operating systems and third party software (non-operating system related) in the manner prescribed by the developer, and then add third party software such as Microsoft’s Office 97 or Office 2000, Corel’s WordPerfect, Adobe Photoshop just the way any user would in their own individual environments. Then, and only then, do we run our performance tests. This enables us to produce validated, reproducible, data based upon actual user environments as opposed to data derived from a computer that has been constructed to do nothing more than perform well in a single test environment. We do not perform any special hardware or software tweaks other than those recommended by the hardware and software manufacturers, all which are published right here on our Web site. We do not add any secret software or use any special devices to enhance performance.
Most performance issues can be reduced to a few factors, mating the right motherboard, processor, memory and hard drives in order to provide the desired result. Usually, the utilization of cheap no-name components will result in poor documentation, poor or non-existent technical support, no warranty replacements and most of all, poor performance.
The Software we use:
- SYSmarkÂ® 2000 by BAPCo: The SYSmarkÂ® 2000 software suites are used to measure real-world application performance. To read more about the testing and applications involved in this suite, click here.
- 3DMarkÂ™ 2000: 3DMark 2000 was designed specfically for DirectX 7, and is the latest of the 3DMark series of gamer’s benchmarks. It tests all the latest updates in 3D gaming hardware and includes a number of graphically intense demo scenes. To read more about 3DMarkÂ™ 2000 click here.
- Video 2000: Video 2000, from the same developers as SYSmark 2000 and 3DMarkÂ™ 2000, have introduced one of the first ever Video performance testing tools that can, among other things, verify how well a computer can play DVD movies and verify the quality of a video image quality. To read more about Video 2000 click here.
- Quake III Arena Demo: Quake III Arena is one of the most graphically intense, (if not the most graphically intense), games on the market today. While there are no true measurements to be made in Quake III Arena Demo, the mere fact that it is one of the most demanding 3D games currently available dictates that any hardware bottlenecks or incorrect configurations will surely result in an instant system performance degradation or failure.
- SiSoft Sandra Professional: While SiSoft Sandra Professional is a synthetic benchmark tool, its diagnostic tools can be used to measure CPU, FPU, memory and multimedia performance to identify and correct certain performance issues. To read more about SiSoft Sandra Professional, click here.
Why do we do it this way?
This is a question that is both simple and difficult to answer. Over the last several years millions of desktop computers, work stations and servers have been sold around the world, and thousands more are being sold every day. If you have spent any time at all looking at the various brands, each manufacturer states rather emphatically that their product is better than all of the others. Well today that might be true, but tomorrow it may not be. Technology is changing so rapidly that it’s difficult enough keeping up with software updates and upgrades, let alone trying to determine from one day to the next who the current leader is in computer development.
When you read all of the advertisements and media hype produced by the leading computer manufacturers, there’s one component missing from each of their design concepts, you! Think about it, do any of them ask you how you intend to use the product you purchase from them? Branded products are great as long as we, the purchaser, don’t get carried away with the theme. There is no such thing as a “one size fits all” in computers.
By standardizing on our testing software and procedures, we are able to provide you, the purchaser, with a close-up view of what the right package of components can do for you to make what you want to do easier, and your end product better. Thousands of computers are sold every day, some of these are cheap (not to be equated with inexpensive) junk, constructed of no-name components. Others are well designed and well manufactured pieces of equipment. It’s up to you to decide whether you want to invest in a computer designed around you, or a computer designed around a concept developed by the manufacturer.
There is a right way and a wrong way to do everything, and the direction you choose will decide how your computer performs for you. Remember, you do get what you pay for!
To read more about Application Based Benchmark Software and The Business Applications Performance Corporation, (BAPCo) please click on this link.
SYSmark 2000, 3DMark 2000, Video 2000 Pro, Sandra and Quake III Arena are registered trademarks or trademarks of their respective owners, in the U.S. and other countries. Our testing was performed following the prescribed methodology described by the respective bench mark software developers. The benchmark developers makes no representations or warranties as to the results of these tests.
. All rights reserved.