https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106440
Bug ID: 106440
Summary: on Mac M1 chip undefined behavior of double arithmetic
Product: gcc
Version: 12.1.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: target
Assignee: unassigned at gcc dot gnu.org
Reporter: yaobig97 at gmail dot com
Target Milestone: ---
Created attachment 53351
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53351&action=edit
the preprocessed file (*.i*) that triggers the bug
Here are the infos you want for reporting bugs:
%% gcc-12 -v %%
Using built-in specs.
COLLECT_GCC=gcc-12
COLLECT_LTO_WRAPPER=/opt/homebrew/Cellar/gcc@12/12.1.0_1/bin/../libexec/gcc/aarch64-apple-darwin21/12/lto-wrapper
Target: aarch64-apple-darwin21
Configured with: ../configure --prefix=/opt/homebrew/opt/gcc@12
--libdir=/opt/homebrew/opt/gcc@12/lib/gcc/12 --disable-nls
--enable-checking=release --with-gcc-major-version-only
--enable-languages=c,c++,objc,obj-c++,fortran --program-suffix=-12
--with-gmp=/opt/homebrew/opt/gmp --with-mpfr=/opt/homebrew/opt/mpfr
--with-mpc=/opt/homebrew/opt/libmpc --with-isl=/opt/homebrew/opt/isl
--with-zstd=/opt/homebrew/opt/zstd --with-pkgversion='Homebrew GCC 12.1.0_1'
--with-bugurl=https://github.com/Homebrew/homebrew-core/issues
--with-system-zlib --build=aarch64-apple-darwin21
--with-native-system-header-dir=/usr/include
--with-sysroot=/Library/Developer/CommandLineTools/SDKs/MacOSX12.sdk
Thread model: posix
Supported LTO compression algorithms: zlib zstd
gcc version 12.1.0 (Homebrew GCC 12.1.0_1)
%% the complete command line that triggers the bug %%
g++-12 -O2 -Wall -Wextra test.cpp
%% the compiler output (error messages, warnings, etc.) %%
No error.
%% the preprocessed file (*.i*) that triggers the bug %%
See the attachment.
Here are the infos which are said not needed but pretty helpful from my side:
%% code %%
#include <iostream>
#include <vector>
using namespace std;
struct P {
double x, y;
P operator -(const P &a) const { return P{x - a.x, y - a.y}; }
double cross(const P &b) const { return x * b.y - y * b.x; }
};
int main() {
const double A = 3'000'000'000;
vector<P> as{P{A, 1 + A}, P{0, 1}};
auto c = as[0];
auto b = as[1];
cout << c.x << " " << c.y << endl;
cout << (P{1, 2} - as[0]).cross(as[1] - as[0]) << endl;
cout << (P{1, 2} - c).cross(b - c) << endl;
return 0;
}
%% output %%
3e+09 3e+09
512
0
%% Description %%
I do not see why the second line and the third line in the output can be
different.
I do not think that the code itself is wrong but the behavior really seems to
be undefined: remove cout << c.x << " " << c.y << endl; and output numbers are
both 0.
The code works perfectly under gcc on x86 and also clang on ARM (Mac M1 chip)
so I suppose that the error comes from gcc on ARM (Mac M1 chip)
