Hacker News new | ask | show | jobs
by bmvakili 4475 days ago
I got widely different results there: C - 105.868545% C++ - 80.0518% Java - 61.664313124608775% Probably the optimizations there; can't easily be done in Java one; interesting nonetheless, thanks for post.
2 comments

Those benchmark ratios are not comparable across languages.

The C code compares running time with a very standard mergesort.

The C++ code compares with std::stable_sort.

The Java code compares with a standard mergesort -- very similar to the code in the C version -- but has hard-to-predict JIT warmup effects.

What are your result numbers supposed to mean?