On Fri, 22 Dec 2023 15:08:33 GMT, Markus KARG <d...@openjdk.org> wrote:
> IMHO the trigger for this PR was sparing time That is fair. Do you know how one could create a reliable test for performance that would fail without the fix and succeed with the fix? I think the allocation is a reasonable proxy for the duration in this case, and the allocation is more-or-less easy to measure and test if the JVM supports `getCurrentThreadAllocatedBytes`. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/16879#discussion_r1435200386
