RFR: 8373026: C2 SuperWord and Vector API: vector algorithms test and benchmark [v10]

[Xiaohong Gong]

On Tue, 20 Jan 2026 07:21:06 GMT, Emanuel Peter <epeter at openjdk.org> wrote:

> > > If the Vector API code can be fully intrinsified, we get fast code. But somtimes, the Vector API is horribly slow, much slower than scalar loop performance.
> > > linux_aarch64_server: filterI, scanAddI, reduceAddIFieldsX4 are very slow
> > 
> > 
> > Hi @eme64 , I noticed that these benchmarks have even much worse performance on AAch64. May I ask whether the machine supports SVE/SVE2 feature? Thanks!
> 
> @XiaohongGong thanks for reviewing!
> 
> No, I only have access to NEON machines, unfortunately. I added a comment to the PR description for that just now :)

Got it. Thanks for your input and update! So these result looks reasonable to me as APIs used in those benchmarks are not supported to be intrinsified with NEON only machines.

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

PR Comment: https://git.openjdk.org/jdk/pull/28639#issuecomment-3771452288