Re:Why is it faster to process a sorted array than an unsort...
by ggeorgovassilis on Monday May 20, 2019 @07:38AM
(#19959365)
After running the experiment (openjdk 8, linux) with some variations, I'm convinced it's a JIT optimisation and not a CPU predictive branch feature, eg. increaseing the threshold from 128 to post-16 bit, eg 100000:
if (data[c] >= 128)
[ Parent
| on HN ]
There are no comments on this comment yet.