14 DAY TRIAL // Despite what seems to have become a generalized status quo, Windows Vista is an operating system radically different in comparison to Windows XP. That is why the configuration for performance measurements has also evolved from XP. Well, Intel has posted concomitantly with the general availability of Windows Vista a selection of best practices for benchmarking the operating system. In the absence of an official methodology for benchmarking Windows Vista from Microsoft, you might consider the set of guidelines published over at the Intel Capabilities Forum.
The best practices for Windows Vista benchmarking is divided into four stages. The procedure involves preparing the system, setting Windows Vista up and the recommended settings for accurate performance testing. Intel recommends that all benchmarking be done for Windows Vista Ultimate edition. This is the actual methodology that Intel applies for testing Windows Vista internally.
The process described by Intel starts at the hardware level. In fact, the quality, status and management of the hardware resources can impact the final results of the operating system's performance measurements. As a result, benchmarkers should pay attention to all the hardware resources of the system that will run Vista for the tests.
As a standard of Windows benchmarking, testers have to clean up the Windows Prefetch folder. Deleting all the contents in c:windowsprefetch will get rid of the cached applications that enable the operating system to benefit from extra speed.
However, in the end it is also about how the various areas of Windows Vista are set up. For this, Intel delivers the optimal settings for a benchmarking starting with Windows Aero, and ending with Windows Update. "We recommend running workloads five (5) times in a row after you Process Idle Tasks without reboot in between, and then taking the median score of the five runs," reads one final advice from Intel.