RFR: 8338967: Improve performance for MemorySegment::fill [v10]

Tue Sep 3 10:28:26 UTC 2024

On Tue, 3 Sep 2024 08:49:58 GMT, Francesco Nigro <duke at openjdk.org> wrote:

>> I've tried 
>> 
>> 
>> final long longValue = Byte.toUnsignedLong(value) * 0x0101010101010101L;
>> 
>> 
>> But it had the same performance as explicit bit shifting on M1.
>
> @minborg the ` ELEM_SIZE` is a `Param` field right? Just to be 100% sure of it...

yes it is. But I think the reason we are not seeing any difference is that in the fill benchmarks, we are always using a value of zero. We should keep this trick in mind for later...

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

PR Review Comment: https://git.openjdk.org/jdk/pull/20712#discussion_r1741814268