Hello, I have been following the development of Ruby 2.4's most exciting feature (for me) Open Address Hash Tables In the development process for this new feature, Koichi Sasada went to great lengths to generate benchmarks of the new feature vs ruby trunk ( here is the report)
I think there's some interesting hash benchmarks he's using that aren't on rubybench. Definitely some opportunities to expand hash benchmarking. Particularly the large table benchmarks would be a nice addition. The issue on ruby-lang.org had more discussed benchmarks.
How does rubybench decide what benchmarks to include?