RFR: 8341697: C2: Register allocation inefficiency in tight loop [v6]

[Vladimir Kozlov]

On Mon, 14 Oct 2024 13:42:45 GMT, Quan Anh Mai <qamai at openjdk.org> wrote:

>> Quan Anh Mai has updated the pull request incrementally with one additional commit since the last revision:
>> 
>>   refine comments + typo
>
> Thanks for the source code. That's really interesting, running the benchmark multiple times may give different results, and even when there is a difference in the observed throughputs, the 2 compiled methods are exactly the same. So I think we are running into different quirks here, probably due to the fact that this benchmark saturates the memory bandwidth.

> @merykitty can you run this with regular Java benchmarks (SPECjvm, SPECjbb, Renaissance, DaCapo) to see if they are affected?

We will also run our set of benchmarks to make sure there is no regression. If we see significant regression only in some benchmarks and improvement in others we can set `LoopAwareSpilling` to false in these changes and address regression in following PR.

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

PR Comment: https://git.openjdk.org/jdk/pull/21472#issuecomment-2414657865