https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127241
Bug ID: 127241
Summary: [17 Regression] 4% slowdown of gamess on Zen5 since
r17-3757-gabd40f513e0882
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
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=1241.50.0
there was a 4% exec time slowdown of the gamess SPEC 2006
benchmark when compiled with -Ofast -march=x86-64-v3 -g -flto=128 on a Zen5
machine. I bisected it to r17-3757-gabd40f513e0882.
commit abd40f513e0882c4132367877c27201f12d234d0
Author: Jovan Dmitrovic <[email protected]>
Date: Sat Aug 29 12:12:34 2026 -0600
[PATCH v4 2/3] ivopts: Remove target_reg_cost from reg pressure estimate.
This is a regression against GCC 16. See the comparison
here:
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=1260.50.0&plot.1=1322.50.0&plot.2=1290.50.0&plot.3=1350.50.0&plot.4=1550.50.0&plot.5=1241.50.0&;
Jovan doesn't have a bugzilla account yet, I'll email him.
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)
