https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126915
Bug ID: 126915
Summary: [17 Regression] 13% slowdown of milc on Zen2 since
r17-3220-g5e63fb58f84e3c
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=299.70.0
there was a 13% exec time slowdown of the milc SPEC 2006
benchmark when compiled with -Ofast -march=native -g PGO on an Zen2
machine. I bisected it to r17-3220-g5e63fb58f84e3c.
5e63fb58f84e3ce19cb336710f69094ecbd4b63d is the first 'bad' commit
commit 5e63fb58f84e3ce19cb336710f69094ecbd4b63d
Author: Richard Biener <[email protected]>
Date: Wed Jul 15 15:33:31 2026 +0200
tree-optimization/126028 - vector placement and live lane extracts
This is a regression against GCC 16. See the comparison
here:
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=328.70.0&plot.1=315.70.0&plot.2=336.70.0&plot.3=281.70.0&plot.4=404.70.0&plot.5=422.70.0&plot.6=760.70.0&plot.7=935.70.0&plot.8=1055.70.0&plot.9=1205.70.0&plot.10=1515.70.0&plot.11=299.70.0&;
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)
