RFR: 8303762: Optimize vector slice operation with constant index using VPALIGNR instruction [v2]

[Xiaohong Gong]

On Wed, 13 Aug 2025 07:02:34 GMT, Jatin Bhateja <jbhateja at openjdk.org> wrote:

>> test/micro/org/openjdk/bench/jdk/incubator/vector/VectorSliceBenchmark.java line 36:
>> 
>>> 34: @State(Scope.Thread)
>>> 35: @Fork(jvmArgs = {"--add-modules=jdk.incubator.vector"})
>>> 36: public class VectorSliceBenchmark {
>> 
>> I remember that it has the micro benchmarks for slice/unslice under `test/micro/org/openjdk/bench/jdk/incubator/vector/operation` on panama-vector. Can we reuse those JMHs to check the benchmark improvement?
>
>> I remember that it has the micro benchmarks for slice/unslice under `test/micro/org/openjdk/bench/jdk/incubator/vector/operation` on panama-vector. Can we reuse those JMHs to check the benchmark improvement?
> 
> All those are the ones with variable slice index , slice kernel performance of those benchmarks on AVX2 and AVX512 targets are at par with baseline, and deviations are statistically insignificant due to error margins. 
> 
> New benchmark complements the code.

OK. Make sense to me. Thanks!

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

PR Review Comment: https://git.openjdk.org/jdk/pull/24104#discussion_r2272278442