I'm with MiRai. To further that, the first round of "official" benchmarks should be taken as incomplete data on general principle as it usually takes a few months of various people running them before we find out which synthetic benchmarks favor which processors and where the lurking demons (e.g. bugs) lie.