Skip to content

Implement the benchmark using Java arrays

This is the Java array based implementation with some JVM options applied on top of the default ones.

I didn't try to generalize the code, so it should be comparable to the HashMap/go implementations pretty directly. The biggest slowdown in the previous G1-based Java HashMap implementation was due to the unspecified Xmx (max heap size) parameter. Having it bounded reduced the longest pause on my machine to ~30ms with HashMap (1gb max heap) and ~15ms with arrays (512m max heap).

Merge request reports

Loading