Reproducible performance

[gitboy16]

Hi,

The way you benchmark the sorting algorithms obviously depends on the hardware you are using.
From a previous issue I opened you can see that depending on the hardware Julia is not the fastest, nor the most memory efficient solution.

I don't have a solution to that but it would be great to find a way for people to be able to reproduce the perfomance that you are displaying.

I was wondering if it is possible to run these benchmarks on a github server or equivalent enviornnment where people can actually reproduce things.

Thank you
Kind regards

[LilithHafner]

Unfortunately, I do not have the bandwidth to configure 8 different languages on a remote server. Feel free to submit a PR, though!

[dumblob]

I think it would be enough to get rid of numeric sorting and suddenly the performance should be much more reproducible if not "always everywhere".

See #11 .

[LilithHafner]

Reproducibility is separate from having a diverse set of benchmarks—though both are important and lacking here.