On Fri, 27 Nov 2020 20:36:55 GMT, Claes Redestad <redes...@openjdk.org> wrote:
>> Done
>
> Nice!
>
> Did this help the non-latin1 case in your microbenchmark?
Yes, now it's significantly faster on longer strings
(latin) (length) Score Error Units
insert true 8 22.970 ± 0.291 ns/op
insert true 64 37.636 ± 0.613 ns/op
insert true 128 45.412 ± 0.876 ns/op
insert true 1024 134.511 ± 16.446 ns/op
insert false 8 35.429 ± 2.410 ns/op
insert false 64 60.654 ± 1.673 ns/op
insert false 128 80.743 ± 4.798 ns/op
insert false 1024 368.987 ± 19.309 ns/op
-------------
PR: https://git.openjdk.java.net/jdk/pull/402
