> This micro-benchmark makes sense. It's benchmarking an HTTP server essentially.

Kind-of true. But to give any results that can be applied to the real world, it would have to use a much larger payload than a hello world.