[Bug tree-optimization/121037] New: [16 Regression] 4-6% slowdown of 482.sphinx3 since r16-2088-ge9079e4f43d135

[pheeck at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=121037

            Bug ID: 121037
           Summary: [16 Regression] 4-6% slowdown of 482.sphinx3 since
                    r16-2088-ge9079e4f43d135
           Product: gcc
           Version: 16.0
            Status: UNCONFIRMED
          Keywords: missed-optimization
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: pheeck at gcc dot gnu.org
                CC: rguenth at gcc dot gnu.org
            Blocks: 26163
  Target Milestone: ---
              Host: x86_64-pc-linux-gnu,
            Target: x86_64-pc-linux-gnu

As seen here

https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=992.280.0
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=965.280.0

there was a 4-6% (i got the 6% when I remeasured this manually) exec time
slowdown of the 482.sphinx3 SPEC 2006
benchmark when run with -Ofast -march=native (-flto) on an AMD Zen4 machine.
I bisected it to r16-2088-ge9079e4f43d135 (2025-07-08).

e9079e4f43d13579c41110ce1871051a43c577b6 is the first bad commit
commit e9079e4f43d13579c41110ce1871051a43c577b6
Author: Richard Biener <rguent...@suse.de>
Date:   Sun May 25 19:29:04 2025 +0200

    add masked-epilogue tuning


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)