Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

You can go back and read the mailing list archives to see the benchmarks that were used by Python devs to decide on using Timsort: https://mail.python.org/pipermail/python-dev/2002-July/threa...

You can also read Tim Peter's notes about the sort: https://bugs.python.org/file4451/timsort.txt

That was back in 2002, so there's certainly been a lot of scrutiny since then. In particular, I've never actually read about Java sorting benchmarks, but there are tremendously sophisticated benchmarking tools/benchmarking experts in the Java world, so I imagine that it got a lot of analysis there.



It's worth noting that Java doesnt exclusively use Timsort. Depeding on the data type and heuristics based mostly on the size of the input array, a simpler sort such as counting sort may be use.

You can find a more detailed summary at:

https://stackoverflow.com/a/41129231/149138


Cool find. I'd never actually realized that.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: