When I try compile code with 128 bit long double arithmetic, I get a request to
file a bug report (and the compile fails). Below is a command line, the source
file, the g++ output. They are separated by "%% <identifier> %%". If you need
the intermediate files test.ii and test.s, I can send them. Without the flag
-mlong-double-128, the code compiles and runs.
%% command line %%
g++-3.4.3 -v -save-temps -mlong-double-128 test.cxx -o test >& g++.out
%% test.cxx %%
#include <iostream>
using namespace std;
int main() {
cout << "sizeof(float)=" << sizeof(float) << endl;
cout << "sizeof(double)=" << sizeof(double) << endl;
cout << "sizeof(long double)=" << sizeof(long double) << endl;
long double a, b;
a = b = 1.0L;
do {
b = b/2;
} while ((a+b)!=a);
cout << "long double epsilon = " << b << endl;
return 0;
}
%% g++ output %%
Reading specs from
/opt/local/lib/gcc-3.4.3/gcc/powerpc-apple-darwin7.8.0/3.4.3/specs
Configured with:
/Users/brian/darwinports/dports/lang/gcc34/work/gcc-3.4.3/configure
--prefix=/opt/local --enable-languages=c,c++,f77
--libdir=/opt/local/lib/gcc-3.4.3 --infodir=/opt/local/share/info
--mandir=/opt/local/share/man --with-system-zlib --disable-nls
--program-suffix=-3.4.3
--with-gxx-include-dir=/opt/local/include/gcc-3.4.3/c++/3.4.3
Thread model: posix
gcc version 3.4.3
/opt/local/libexec/gcc/powerpc-apple-darwin7.8.0/3.4.3/cc1plus -E -quiet -v
-D__DYNAMIC__ test.cxx -fPIC -mlong-double-128 -o test.ii
ignoring nonexistent directory
"/opt/local/lib/gcc-3.4.3/gcc/powerpc-apple-darwin7.8.0/3.4.3/../../../../powerpc-apple-darwin7.8.0/include"
#include "..." search starts here:
#include <...> search starts here:
/opt/local/include/gcc-3.4.3/c++/3.4.3
/opt/local/include/gcc-3.4.3/c++/3.4.3/powerpc-apple-darwin7.8.0
/opt/local/include/gcc-3.4.3/c++/3.4.3/backward
/usr/local/include
/opt/local/include
/opt/local/lib/gcc-3.4.3/gcc/powerpc-apple-darwin7.8.0/3.4.3/include
/usr/include
End of search list.
/opt/local/libexec/gcc/powerpc-apple-darwin7.8.0/3.4.3/cc1plus -fpreprocessed
test.ii -fPIC -quiet -dumpbase test.cxx -mlong-double-128 -auxbase test -version
-o test.s
GNU C++ version 3.4.3 (powerpc-apple-darwin7.8.0)
compiled by GNU C version 3.4.3.
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
test.cxx: In function `int main()':
test.cxx:16: error: insn does not satisfy its constraints:
(insn 88 165 166 (set (reg:DI 35 f3)
(mem/u/f:DI (lo_sum:SI (reg:SI 2 r2)
(const:SI (minus:SI (symbol_ref/u:SI ("*LC4") [flags 0x2])
(symbol_ref:SI ("<pic base>"))))) [0 S8 A64])) 327
{*movdi_internal32} (nil)
(nil))
test.cxx:16: internal compiler error: in final_scan_insn, at final.c:2359
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.
--
Summary: -mlong-double-128 arithmetic error
Product: gcc
Version: 3.4.3
Status: UNCONFIRMED
Severity: normal
Priority: P2
Component: c++
AssignedTo: unassigned at gcc dot gnu dot org
ReportedBy: coomes at math dot miami dot edu
CC: coomes at math dot miami dot edu,gcc-bugs at gcc dot gnu
dot org
GCC build triplet: powerpc-apple-darwin7.8.0
GCC host triplet: powerpc-apple-darwin7.8.0
GCC target triplet: powerpc-apple-darwin7.8.0
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=20820
