https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125572
Bug ID: 125572
Summary: [17 Regression] 4% slowdown of milc on Zen2 since
r17-837-g8af2e8e49d6e5d
Product: gcc
Version: 17.0
Status: UNCONFIRMED
Keywords: missed-optimization
Severity: normal
Priority: P3
Component: tree-optimization
Assignee: unassigned at gcc dot gnu.org
Reporter: jmelcr at gcc dot gnu.org
CC: tnfchris at gcc dot gnu.org
Blocks: 26163
Target Milestone: ---
As seen here
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=291.70.0
there was a 4% exec time slowdown of the milc SPEC 2006
benchmark when compiled with -O2 -march=native -g on a Zen2
machine. I bisected it to r17-837-g8af2e8e49d6e5d.
8af2e8e49d6e5d33c01c2beaead4933bc286974c is the first bad commit
commit 8af2e8e49d6e5d33c01c2beaead4933bc286974c
Author: Tamar Christina <[email protected]>
Date: Wed May 27 10:53:07 2026 +0100
vect: Don't generate scalar epilogue if not needed [PR120352]
This is a regression against GCC 15, on par with GCC 16. See the comparison
here:
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=332.70.0&plot.1=304.70.0&plot.2=340.70.0&plot.3=272.70.0&plot.4=392.70.0&plot.5=411.70.0&plot.6=756.70.0&plot.7=925.70.0&plot.8=1051.70.0&plot.9=1195.70.0&plot.10=1501.70.0&plot.11=291.70.0&;
Referenced Bugs:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=26163
[Bug 26163] [meta-bug] missed optimization in SPEC (2026, 2k17, 2k and 2k6 and
95)
