I do not usually believe in any vendor provided performance benchmark unless I fabricated it myself. On a more serious note, benchmarking is pretty hard and peo

I do not usually believe in any vendor provided performance benchmark unless I fabricated it myself. On a more serious note, benchmarking is pretty hard and people usually discount things that seem minor until you found it the bottleneck, like the interconnect in this case for example. Another problem with synthetic benchmarks is that you can always optimize for your exact use case and it usually yields to pretty good improvements, comparable to buying faster equipment. The ultimate quiestion which is more cost efficient, buying faster CPU/GPU or hiring a performance expert.