https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118967
Bug ID: 118967
Summary: 5% slowdown of 473.astar on AMD Zen3 since
r15-7400-gd3ff498c478ace
Product: gcc
Version: 15.0
Status: UNCONFIRMED
Keywords: missed-optimization, ra
Severity: normal
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: pheeck at gcc dot gnu.org
CC: hjl at gcc dot gnu.org
Blocks: 26163
Target Milestone: ---
Host: x86_64-linux
Target: x86_64-linux
As seen here
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=464.260.0
there was a 5% exec time slowdown of the 473.astar SPEC 2006 benchmark caused
by commit r15-7400-gd3ff498c478ace
when run with -O2 -march=native on an AMD Zen3 machine.
Cc-ing H. J. Lu and Honza.
I have looked at more astar results. There are some configurations where astar
speeds up a bit. Here is one example, but it is quite noisy:
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=292.260.0
However, the slowdowns are more numerous and bigger.
---
This is *not* a regression against GCC14. See the comparison
here:
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.8=1037.260.0&plot.9=464.260.0&;
Referenced Bugs:
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=26163
[Bug 26163] [meta-bug] missed optimization in SPEC (2k17, 2k and 2k6 and 95)
