https://gcc.gnu.org/bugzilla/show_bug.cgi?id=127400
Bug ID: 127400
Summary: [17 Regression] 7-9% slowdown of calculix on aarch64
since r17-3895-gc9698b7443d6f1
Product: gcc
Version: 17.0
Status: UNCONFIRMED
Keywords: missed-optimization
Severity: normal
Priority: P3
Component: target
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: aarch64-suse-linux
Target: aarch64-suse-linux
Build: aarch64-suse-linux
Hi,
we are seeing several calculix slowdowns on Aarch64 (ARM Neoverse N1):
7% slowdown with -O2 -g -flto=128:
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=574.170.0
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=1035.170.0&plot.1=1220.170.0&plot.2=1506.170.0&plot.3=574.170.0&;
9% with -O2 -g -flto=128 PGO:
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=686.170.0
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=1049.170.0&plot.1=1223.170.0&plot.2=1519.170.0&plot.3=686.170.0&;
The slowdowns bisect to r17-3895-gc9698b7443d6f1:
commit c9698b7443d6f19a7bac86e34d6eaeb5439fff60
Author: Richard Biener <[email protected]>
Date: Wed Sep 2 13:05:05 2026 +0200
[x86] move CTOR gpr<->xmm move costing to add_slp_cost overload
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)
