ICE with gcc-trunk-x86_64 20070508 on aermod (polyhedron test)
Dominique Dhumieres
dominiq@lps.ens.fr
Wed May 9 08:11:00 GMT 2007
(gdb) run -w -O3 -ffast-math -funroll-loops aermod.f90
Starting program: /users/stat5/dominiq/SUN/test/gfortran/gcc-trunk/libexec/gcc/x86_64-unknown-linux-gnu/4.3.0/f951 -w -O3 -ffast-math -funro
ll-loops aermod.f90
MAIN__ hrloop julian gregor hrqext o3ext
Program received signal SIGSEGV, Segmentation fault.
0x0000002a958c29bd in __gmpz_sub () from /users/stat5/dominiq/SUN/test/gfortran/gcc-trunk/lib64/libgmp.so.3
(gdb) backtrace
#0 0x0000002a958c29bd in __gmpz_sub () from /users/stat5/dominiq/SUN/test/gfortran/gcc-trunk/lib64/libgmp.so.3
#1 0x000000000044d7d2 in resolve_ref (expr=0x16112a0) at /home/tob/projects/gcc/gcc/fortran/resolve.c:2906
#2 0x000000000044ded7 in gfc_resolve_expr (e=0x16112a0) at /home/tob/projects/gcc/gcc/fortran/resolve.c:3130
#3 0x00000000004519a9 in resolve_code (code=0x1611630, ns=0x158ceb0) at /home/tob/projects/gcc/gcc/fortran/resolve.c:5107
#4 0x000000000045158e in gfc_resolve_blocks (b=0x1611af0, ns=0x158ceb0) at /home/tob/projects/gcc/gcc/fortran/resolve.c:5040
#5 0x0000000000451989 in resolve_code (code=0x1611bf0, ns=0x158ceb0) at /home/tob/projects/gcc/gcc/fortran/resolve.c:5099
#6 0x000000000045314e in resolve_codes (ns=0x158ceb0) at /home/tob/projects/gcc/gcc/fortran/resolve.c:7407
#7 0x0000000000453185 in gfc_resolve (ns=0x158ceb0) at /home/tob/projects/gcc/gcc/fortran/resolve.c:7426
#8 0x00000000004476d8 in gfc_parse_file () at /home/tob/projects/gcc/gcc/fortran/parse.c:3248
#9 0x000000000046646e in gfc_be_parse_file (set_yydebug=Variable "set_yydebug" is not available.
) at /home/tob/projects/gcc/gcc/fortran/f95-lang.c:305
#10 0x0000000000681324 in toplev_main (argc=Variable "argc" is not available.
) at /home/tob/projects/gcc/gcc/toplev.c:1051
#11 0x0000002a95a2100d in __libc_start_main () from /lib64/tls/libc.so.6
#12 0x0000000000405599 in _start ()
#13 0x0000007fbffff298 in ?? ()
...
So the culprit seems the gmp library which has already caused me some trouble
in the past, however it is the same I have used in the past weeks without
problem.
I'll try to have a better look at that later.
Thanks
Dominique
More information about the Fortran
mailing list