https://gcc.gnu.org/bugzilla/show_bug.cgi?id=116758
Bug ID: 116758
Summary: [15 Regression] 25-40% binary size increase and up to
177% compile time increase for SPEC CPU wrf with Ofast
Product: gcc
Version: 15.0
Status: UNCONFIRMED
Keywords: compile-time-hog
Severity: normal
Priority: P3
Component: other
Assignee: unassigned at gcc dot gnu.org
Reporter: pheeck at gcc dot gnu.org
Blocks: 26163
Target Milestone: ---
Host: x86_64-pc-linux-gnu
Target: x86_64-pc-linux-gnu
Between
r15-3465-gde3ca363811a39
r15-3594-g5958279509c460
there was a significant increase in binary size and compile time of the 2017
wrf_r SPEC CPU benchmark and also of its 2006 version.
The regressions seem to happen only on Ofast. I've seen them on our x86_64
machines. I didn't see them on our aarch64 machine. So maybe this is
x86_64-specific.
Here are some examples:
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=959.548.8
521.wrf_r AMD zen4 -Ofast -march=native -flto
156% compilation time increase
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=959.539.4
521.wrf_r AMD zen4 -Ofast -march=native -flto
47% binary size increase
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=1119.270.8
481.wrf AMD zen4 -Ofast -march=x86-64-v3 -flto
112% compilation time increase
https://lnt.opensuse.org/db_default/v4/SPEC/graph?plot.0=793.548.8
521.wrf_r Intel Ice Lake (3rd generation Xeon) -Ofast -march=generic
80% compilation time increase
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)