Multitasking anyone? What happened to multitask testing? How does a given system compare to another if certain applications might be running on their fullest in the background? Let's say DivX and XviD do not seem that good in multithread. So give them a single thread to work with, and play a game at the same time. All those measurements of power efficiency and yet no efficient usage of time really benchmarked somehow. Aren't software always improving on user interaction, especially on the "without" user interaction part, like a batch job. So, give the 4 system 4 consecutive video conversions, or 4 parallel ones. Apart from the first being harder to do, software advancements or not, I believe the second option works best. Still, seeing it benchmarked would help. Like cache usage and cache pollution, bus bottleneck, and of course HDD, but many do have different HDDs for different usages, so a game and a winrar compression should be only bottlenecked from cache and bus.