RFR: 8323116: [REDO] Computational test more than 2x slower when AVX instructions are used [v5]

[Srinivas Vamsi Parasa]

On Fri, 5 Apr 2024 18:17:00 GMT, Sandhya Viswanathan <sviswanathan at openjdk.org> wrote:

>> My new testing passed. 
>> But I want to hear an answer to @merykitty suggestion about using xmm15.
>
> @vnkozlov  If I understand the proposal from @merykitty correctly, the suggestion is to reserve xmm15 as non allocatable throughout. This sounds like a big overhead for cases where every xmm register is usable say in a Vector API kernel. From Vamsi's microbenchmark runs, he has clearly shown that the gain of his optimization is way more than any overhead of doing pxor just before the converts.

> Okay. I will wait changes @sviswa7 suggested to use vlRegD and vlRegF.

Please see the updated commit which uses vlRegD and vlRegF.

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

PR Comment: https://git.openjdk.org/jdk/pull/18503#issuecomment-2040529692