Since much of the implementation is in available libraries, this is a comparison of those too, which is an important consideration. There are language features that slow things down too. The benchmark game makes it clear that no language with GC gets closer than a 2-3x the time of a non-GC language (julia is the closest) and that matches most real world experience (and also that we don't always care about that hit). It also makes it clear that being dynamically typed and interpreted is a huge penalty in performance. I.e. python. But again, development speed can be more important than that throughput hit. Benchmarks can help us decide which tools we like for which tasks.