Conclusion
If you manage to work your way through all the benchmarks we've described, you will definitely learn firsthand about the effort and time it takes for extensive benchmark tests.
We also recommend you take some time to look at a few of our special benchmark articles. Throughout it all, be sure to keep in mind what kinds of results you want your benchmarking quest to yield.
Results will vary somewhat depending on the particular benchmark you use, for example. In some cases the results of individual benchmark cycles fluctuate so widely that the results become meaningless.
In any event, you should never be content with a single result; you should always run a benchmark at least two or three times and take the maximum or average value as the result. Caution is likewise advised when dealing with systems that use Hyperthreading since this feature invariably impacts the odd benchmark result- usually positively, sometimes negatively.
The list of factors that can affect benchmark results is long. Thus, various driver versions, BIOS settings, applications and services running in the background as well as different hardware components (audio, network, USB, FireWire and RAID components) can all exert influences that are difficult to predict. Paying close attention to all of these details and documenting them in case of doubt are required to achieve optimum accuracy and thoroughness. This is the only way to achieve benchmark results that are meaningful and above all, replicable.
