Article
TL;DR: Microsoft reconfigured the RTC in Windows 8 for low-end devices and it causes the RTC to be off when the BCLK is adjusted. Benchmarking programs utilize the RTC to measure performance over time. If the RTC reports 5 minutes, but the actual run time was 5 minutes and 12 seconds, the benchmark would report higher performance than it should.
Interesting read. I realize this isn't the sole reason for the claim, but I'm curious as to how much this has to play in with Windows 8 being "faster" for gaming, etc.
Add: It happens when the BCLK is adjusted after the computer has booted into Windows (ie: via software).