On Tue, Jan 7, 2020 at 12:36 AM Ralph Goers <ralph.go...@dslextreme.com> wrote: > I am not against making a change but I would need to see performance results > before it gets merged.
Unless my knowledge on the JVM is fundamentally wrong, no shared data structure can beat a TLA in terms of access costs. But at which point synchronization costs of a shared data structure outweigh its advantage -- predictable (and user configurable) memory footprint? I don't think anybody would be willing to realize such a change without convincing benchmark figures, hence agreeing with your comment.