Not sure what's worse here - people relying on third party benchmarks (hint: always do your own; see how a tool performs on your data, on your hardware, for your problem set), or the fanboy-ish panic when they are unsettled that a benchmark might make their chosen toy less shiny?