https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126342
Bug ID: 126342
Summary: [17 Regression] 6% slowdown of deepsjeng_r on Zen2
since r17-1845-g23b6a9a41a5c05
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: konstantinos.eleftheriou at vrull dot eu,
ptomsich at gcc dot gnu.org
Blocks: 26163
Target Milestone: ---
Host: x86_64-suse-linux
Target: x86_64-suse-linux
Build: x86_64-suse-linux
As seen here
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=293.387.0
there was a 6% exec time slowdown of the deepsjeng_r SPEC 2017
benchmark when compiled with -Ofast -march=native -g with PGO on an Zen2
machine. I bisected it to r17-1845-g23b6a9a41a5c05.
commit 23b6a9a41a5c052f880f26b896357c5b207ca24a
Author: Konstantinos Eleftheriou <[email protected]>
Date: Thu Jun 18 09:48:05 2026 +0200
tail-merge: Combine conditions of merged blocks for ccmp [PR102793]
This is a regression against GCC 16. See the comparison
here:
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=333.387.0&plot.1=319.387.0&plot.2=344.387.0&plot.3=273.387.0&plot.4=396.387.0&plot.5=425.387.0&plot.6=783.387.0&plot.7=931.387.0&plot.8=1075.387.0&plot.9=1217.387.0&plot.10=1533.387.0&plot.11=293.387.0&;
There were also these <benchmark> slowdowns in the same timeframe (so probably
caused by the same commit):
4% Zen3 -Ofast -march=native -g with PGO
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=473.387.0
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)
