https://gcc.gnu.org/bugzilla/show_bug.cgi?id=124609
Bug ID: 124609
Summary: [16 Regression] 5-8% slowdown of mcf on x86 since
r16-8128-g5f1024922d3f67 with PGO
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: jmelcr at gcc dot gnu.org
CC: rguenth at gcc dot gnu.org
Blocks: 26163
Target Milestone: ---
As seen here
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=1109.60.0
there was a 5-8% exec time slowdown of the mcf SPEC 2006
benchmark when compiled with -O2 -march=x86-64-v3 -flto=128 with PGO on a Zen 4
machine. I bisected it to r16-8128-g5f1024922d3f67.
5f1024922d3f675852196cdd45b47935892c2c96 is the first bad commit
commit 5f1024922d3f675852196cdd45b47935892c2c96
Author: Richard Biener <[email protected]>
Date: Fri Mar 13 13:11:56 2026 +0100
tree-optimization/122380 - avoid PRE/hoisting of incompatible access-path
This is a regression against GCC 15. See the comparison
here:
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=1111.60.0&plot.1=1130.60.0&plot.2=1121.60.0&plot.3=1136.60.0&plot.4=1147.60.0&plot.5=1109.60.0&;
There were also these mcf slowdowns in the same timeframe (so probably
caused by the same commit):
5% Zen5 -Ofast -march=native with PGO
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=1247.60.0
6% Intel Ice Lake (3rd generation Xeon) -Ofast -march=native with PGO
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=791.60.0
5% Zen3 -Ofast -march=native with PGO
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=468.60.0
4% Zen4 -Ofast -march=native with PGO
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=962.60.0
6% Zen5 -O2 -march=x86-64-v3 -flto=128 with PGO
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=1289.60.0
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)
