RFR: 8297092: [macos_aarch64] Add support for SHA feature detection

[Hao Sun]

On Mon, 28 Nov 2022 10:27:18 GMT, Andrew Haley <aph at openjdk.org> wrote:

> On 11/28/22 02:20, Hao Sun wrote: Note that UseSHA3Intrinsics are NOT enabled by default on supported hardware because there is performance regression one some hardware. See JDK-8292894. Hence we have to enable it explicitly
> I looked for this. On which platforms does performance regress? And by how much?

The performance regression was discussed with @dgbo in [JDK-8295698](https://bugs.openjdk.org/browse/JDK-8295698) (See the comment https://bugs.openjdk.org/browse/JDK-8295698?focusedCommentId=14532738&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-14532738 and after)

In my local evaluation, the performance regression occurred on Graviton3 (Neoverse V1) and Neoverse N2.
The regression is about 25% ~ 30%.

As @dgbo stated, 

> IMHO, the performance benifit of SHA3 intrinsics depends on the micro architecture, it should be switched on/off based on the running platform.

> On 11/28/22 02:53, Hao Sun wrote: 1) should we enable `UseSHA3Intrinsics` by default on M1?
> Yes.
> 2) if so, should we do this in current patch, or one separate one, e.g., JDK-8292894?
> Here is fine.

Thanks for your reply. I will update it soon.

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

PR: https://git.openjdk.org/jdk/pull/11382