https://gcc.gnu.org/bugzilla/show_bug.cgi?id=125931
Bug ID: 125931
Summary: [17 Regression] 6% slowdown of gamess on Zen2 since
r17-1655-g7d351b07e5b85c
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: rguenth 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=289.50.0
there was a 6% exec time slowdown of the gamess SPEC 2006
benchmark when compiled with -Ofast -march=native -g -flto=128 on an Zen2
machine. I bisected it to r17-1655-g7d351b07e5b85c.
7d351b07e5b85c518b5cbe06a6a57cb653911195 is the first bad commit
commit 7d351b07e5b85c518b5cbe06a6a57cb653911195
Author: Richard Biener <[email protected]>
Date: Wed Jun 17 13:19:56 2026 +0200
Add vec_deconstruct costing kind
This is a regression against previous GCC versions. See the comparison
here:
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=322.50.0&plot.1=307.50.0&plot.2=343.50.0&plot.3=266.50.0&plot.4=395.50.0&plot.5=412.50.0&plot.6=765.50.0&plot.7=927.50.0&plot.8=1036.50.0&plot.9=1194.50.0&plot.10=1489.50.0&plot.11=289.50.0&;
There were also these gamess slowdowns in the same timeframe (so probably
caused by the same commit):
6% Zen3 -Ofast -march=native -g -flto=128
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=476.50.0
5% Zen5 -Ofast -march=native -g -flto=128
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=1284.50.0
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)
