PCMark Bias?

[Talcite]

I didn't see this on DT, so I thought I'd put it up here.

Ars Technica came up with some interesting finds about the PCMark2005 memory benchmark. There's some significant performance discrepancies when the CPUID is changed from CentaurHauls (VIA) to GenuineIntel. Here's the link

Has AT investigated this before?

 

[Chiropteran]

Futuremark's benchmarks have always been fishy. Whether this is due to incompetence or malice either way it's just more proof that they should be avoided.