RFR: 8352075: Perf regression accessing fields [v23]

[Radim Vansa]

On Tue, 10 Jun 2025 10:14:21 GMT, Johan Sjölen <jsjolen at openjdk.org> wrote:

>> What's wrong about `memcpy`, or rather the builtin version? Naturally I could write a for cycle copying the bytes, and rely on the compiler to optimize that out anyway, but I think that this makes the intention clear.
>> 
>> If the handling was done through OS, I guess that the penalty would be actually quite severe. I could have tested the previous version on aarch64 e.g. in AWS, though now there's no casting of pointers anymore.
>> 
>> When we have a final version, I could set up a build in AWS and report performance data from there.
>
>>What's wrong about memcpy, or rather the builtin version?
> 
> Doesn't regular `memcpy` compile into the builtin anyway? Aren't there LE/BE concerns when you do this type of computation?

>From what I read `memcpy` should be treated as builtin but in debugger I've seen deeper stacks.

Anyway, this code is gone, I didn't really think about big endian archs.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/24847#discussion_r2137489958