[Bug fortran/43050] -fopenmp causes segementation fault with large arrays

2010-02-12 Thread rknochenmuss at gmx dot net
--- Comment #4 from rknochenmuss at gmx dot net 2010-02-12 20:22 --- Removing the stack limitation is slightly more complicated than I reported earlier: I have found it is not only necessary to use the link option previously suggested by Dominique, it may also be necessary to set this e

[Bug fortran/43050] -fopenmp causes segementation fault with large arrays

2010-02-12 Thread rknochenmuss at gmx dot net
--- Comment #3 from rknochenmuss at gmx dot net 2010-02-12 17:18 --- Thanks guys. Apparently it is a stack limitation. Using ulimit -a, the default is: stack size (kbytes, -s) 8192 ulimit-s unlimited is not permitted. The largest value allowed is 2, this isnt enough to avoid the seg

[Bug fortran/43050] -fopenmp causes segementation fault with large arrays

2010-02-12 Thread dominiq at lps dot ens dot fr
--- Comment #2 from dominiq at lps dot ens dot fr 2010-02-12 12:53 --- As far as I know the "unlimited" stack size on Darwin cannot be larger than 64Mbytes. If you need/want more you need some special incantations: -Wl,-stack_size,... (see http://gcc.gnu.org/bugzilla/show_bug.cgi?id=421

[Bug fortran/43050] -fopenmp causes segementation fault with large arrays

2010-02-12 Thread burnus at gcc dot gnu dot org
--- Comment #1 from burnus at gcc dot gnu dot org 2010-02-12 12:40 --- I cannot reproduce the problem with the array you have, but I can imagine one problem with the real program. By default, local variables of subroutines/functions are put into STATIC memory; this no longer works if th