There has been plenty of fanfare about the arrival of the latest incarnation of Windows*, called Windows* Vista*. Highlights include a new GUI, new APIs, and new search capabilities. But new features can bring with them new challenges, especially for performance measurement.

Experienced benchmarkers may recall some of the initial issues during the early days of Windows* XP. There were issues involving that OS's features to optimize user experience, and background services possibly introducing variability into test results. Eventually, these issues got ironed out, but with the arrival of Windows* Vista*, we thought we'd share our best known method (a "BKM" in Intel-speak) for benchmarking on Vista*.

So without further ado, here's how we benchmark Windows* Vista* in our labs.