RFR: 8249893: AARCH64: optimize the construction of the value from the bits of the other two

[Boris Ulasevich]

Hi Andrew,

Many thanks for looking at this again!

Benchmark link: [1]. Measurements on Cortex A73, Cortex A53 and Neoverse N1
shows [2] 6-15% performance improvement for bench1 and 18-29% for bench2.

thanks,
Boris

[1] http://cr.openjdk.java.net/~bulasevich/8249893/webrev.02/ConstructFF.java
[2] http://cr.openjdk.java.net/~bulasevich/8249893/webrev.02/ConstructFF.txt

On Thu, Oct 8, 2020 at 5:07 PM Andrew Haley <aph at redhat.com> wrote:
>
> On 05/10/2020 18:40, Boris Ulasevich wrote:
> > Let me revive the change request to C2 and AArch64 that applies Bitfield Insert instruction in the expression "(v1 &
> > 0xFF) | ((v2 & 0xFF) << 8)".
> >
> > Compared to the last round of review [2] I updated the transformation to apply BFI in more cases and added a jtreg test.
>
> I looked through the dicussion and I can't find an updated benchmark which
> shows the speedup for the cases you now handle. Is there one?
>
> --
> Andrew Haley  (he/him)
> Java Platform Lead Engineer
> Red Hat UK Ltd. <https://www.redhat.com>
> https://keybase.io/andrewhaley
> EAC8 43EB D3EF DB98 CC77 2FAD A5CD 6035 332F A671
>