RFR: 8362504: AArch64: Replace MOVZ+MOVK+MOVK with ADRP+ADD

Andrew Haley aph at openjdk.org
Thu Aug 21 09:09:03 UTC 2025


On Wed, 20 Aug 2025 08:44:30 GMT, Andrew Haley <aph at openjdk.org> wrote:

>>> In short, in "production" run we need to know what code we are working on/patching: AOTed or normal JITed.
>> 
>> Or code should be the same.
>
>> > In short, in "production" run we need to know what code we are working on/patching: AOTed or normal JITed.
>> 
>> Or code should be the same.
> 
> OK, so we should not commit this patch. it's not worth it.

> @theRealAph @adinn @vnkozlov
> 
> I agree with your points.
> 
> Thank you for taking the time to review and provide such a detailed explanation. I’ve learned a lot from the discussions.
> 
> I’ll close this PR.

Thank you. I hope you're not too discouraged by this. Performance-related patches are often a judgement call, and we often try something and it doesn't work. There's no other way to find out.

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

PR Comment: https://git.openjdk.org/jdk/pull/26653#issuecomment-3209661085


More information about the hotspot-dev mailing list