HTTP Pipelining should not be the base for a performance benchmark. It's an unrealistic scenario.

To make the benchmark a bit fair, add a graphic of memory usage, pipelining might have a huge impact on resources usage in the server side.