RFR: 8309130: x86_64 AVX512 intrinsics for Arrays.sort methods (int, long, float and double arrays) [v42]
Sandhya Viswanathan
sviswanathan at openjdk.org
Fri Oct 13 23:02:37 UTC 2023
On Fri, 13 Oct 2023 10:31:14 GMT, himichael <duke at openjdk.org> wrote:
>> @himichael Please refer to [this question](https://stackoverflow.com/questions/504103/how-do-i-write-a-correct-micro-benchmark-in-java) for how to correctly benchmark Java code.
>
>> @himichael Please refer to [this question](https://stackoverflow.com/questions/504103/how-do-i-write-a-correct-micro-benchmark-in-java) for how to correctly benchmark Java code.
>
> thanks for your reply, but I think this has nothing to do with benchmark.
> I sort a random array of int[], use: ```java.util.Arrays.sort()```
> in jdk 8, run command:
> ```java JDKSort``` , the result display: 15079 ms
>
> in open jdk 22.19 , run command:
> ```java --add-modules jdk.incubator.vector JDKSort```
> the result display: 11619 ms
>
> my question is that this feature should improve performance several times, but it doesn't look like there's much difference between open jdk 22.19 and jdk 8.
> is there a problem with my configuration ?
@himichael Please try java -Xbatch -XX:-TieredCompilation JDKSort. The method DualPivotQuickSort.sort() needs to be C2 compiled for you to see the benefit.
-------------
PR Comment: https://git.openjdk.org/jdk/pull/14227#issuecomment-1762361069
More information about the hotspot-compiler-dev
mailing list