[Bug target/114656] New: ~5% slowdown of 538.imagick_r on aarch64

[Bug target/114656] New: ~5% slowdown of 538.imagick_r on aarch64

[pheeck at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114656

            Bug ID: 114656
           Summary: ~5% slowdown of 538.imagick_r on aarch64
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Keywords: missed-optimization, needs-bisection
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: pheeck at gcc dot gnu.org
            Blocks: 26163
  Target Milestone: ---
              Host: aarch64-gnu-linux
            Target: aarch64-gnu-linux

As seen here

https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=585.507.0

between commits

r14-9649-gbb04a11418f54c
r14-9728-g6fc84f680d098f

there was a 5% exec time slowdown of 538.imagick_r SPEC2017 benchmark on
aarch64 with compilation options -Ofast

5% slowdown can be also seen with -Ofast -march=native

https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=584.507.0

The CPU is Ampere Altra Neoverse N1.

Btw here are the same graphs including also data for GCC13. It is clear that
this isn't a regression against GCC13.

https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.7=882.507.0&plot.8=585.507.0&
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.7=881.507.0&plot.8=584.507.0&


Referenced Bugs:

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=26163
[Bug 26163] [meta-bug] missed optimization in SPEC (2k17, 2k and 2k6 and 95)

[Bug target/114656] ~5% slowdown of 538.imagick_r on aarch64

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114656

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Can you try to revert r14-9692 if that commit isn't the cause?

[Bug target/114656] ~5% slowdown of 538.imagick_r on aarch64

[pheeck at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114656

--- Comment #2 from Filip Kastl <pheeck at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #1)
> Can you try to revert r14-9692 if that commit isn't the cause?

I have tried reverting r14-9692 and that indeed removed the slowdown. The
benchmark ran as fast as on r14-9649-gbb04a11418f54c. So it seems that r14-9692
is the cause of this slowdown.

[Bug target/114656] [15 Regression] ~5% slowdown of 538.imagick_r on aarch64 since r14-9692

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114656

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |15.0
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2024-04-14
            Summary|~5% slowdown of             |[15 Regression] ~5%
                   |538.imagick_r on aarch64    |slowdown of 538.imagick_r
                   |since r14-9692              |on aarch64 since r14-9692
             Status|UNCONFIRMED                 |NEW

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
. Patch weas reverted but it will be added back after branching so ...