RFR: 8322174: RISC-V: C2 VectorizedHashCode RVV Version [v17]

[Yuri Gaevsky]

On Wed, 6 Aug 2025 10:13:46 GMT, Yuri Gaevsky <duke at openjdk.org> wrote:

>> The patch adds possibility to use RVV instructions for faster vectorizedHashCode calculations on RVV v1.0.0 capable hardware.
>> 
>> Testing: hotspot/jtreg/compiler/ under QEMU-8.1 with RVV v1.0.0.
>
> Yuri Gaevsky has updated the pull request incrementally with one additional commit since the last revision:
> 
>   make 'result' calculations scalar; clear vector registers only when necessary.

`7c5f24a`:

--- -XX:+UseRVV ---
Benchmark            (size)  Mode  Cnt     Score    Error  Units
ArraysHashCode.ints       1  avgt   30    11.277 ±  0.006  ns/op
ArraysHashCode.ints       5  avgt   30    23.300 ±  0.147  ns/op
ArraysHashCode.ints      10  avgt   30    38.838 ±  0.017  ns/op
ArraysHashCode.ints      20  avgt   30    70.163 ±  0.025  ns/op
ArraysHashCode.ints      30  avgt   30   101.495 ±  0.082  ns/op
ArraysHashCode.ints      40  avgt   30    69.880 ±  0.390  ns/op
ArraysHashCode.ints      50  avgt   30   103.447 ±  0.519  ns/op
ArraysHashCode.ints      60  avgt   30   134.626 ±  0.302  ns/op
ArraysHashCode.ints      70  avgt   30    74.650 ±  0.142  ns/op
ArraysHashCode.ints      80  avgt   30   104.745 ±  0.458  ns/op
ArraysHashCode.ints      90  avgt   30   136.548 ±  0.090  ns/op
ArraysHashCode.ints     100  avgt   30    74.526 ±  0.010  ns/op
ArraysHashCode.ints     200  avgt   30   115.240 ±  1.812  ns/op
ArraysHashCode.ints     300  avgt   30   157.611 ±  2.621  ns/op
ArraysHashCode.ints    1000  avgt   30   343.443 ±  9.225  ns/op
ArraysHashCode.ints   10000  avgt   30  5426.387 ± 30.774  ns/op

-------------

PR Comment: https://git.openjdk.org/jdk/pull/17413#issuecomment-3159965574