[mpich-discuss] Problem with Intel Fortran Compiler and MPICH2 on Mac OS X

Omer Acikgoz oacikgoz at hotmail.com
Tue Oct 26 16:20:13 CDT 2010


I forgot to indicate that this is a Dual 6-core Xeon Westmere Mac Pro.

On Oct 26, 2010, at 5:14 PM, Omer Acikgoz wrote:

> I have a problem using Intel Fortran Compiler 11.1 with MPICH2 version 1.3 on Mac OS X.
> 
> I do not have Intel C/C++ compiler, therefore when I compiled MPICH2, the C libraries were compiled using GCC 4.6 and I used ifort as the default fortran compiler. The compilation had no errors or problems, however, when I compiled my first code using mpif90, this is what I got
> 
> Undefined symbols:
>  "___emutls_get_address", referenced from:
>      _MPIR_Allreduce_intra in libmpich.a(allreduce.o)
>      _MPIR_Allreduce_intra in libmpich.a(allreduce.o)
>      _MPIR_MINF in libmpich.a(opmin.o)
>      _MPIR_BAND in libmpich.a(opband.o)
>      _MPIR_MAXLOC in libmpich.a(opmaxloc.o)
>      _MPIR_PROD in libmpich.a(opprod.o)
>      _MPIR_BXOR in libmpich.a(opbxor.o)
>      _MPIR_MINLOC in libmpich.a(opminloc.o)
>      _MPIR_BOR in libmpich.a(opbor.o)
>      _MPIR_LXOR in libmpich.a(oplxor.o)
>      _MPIR_Reduce_intra in libmpich.a(reduce.o)
>      _MPIR_Reduce_intra in libmpich.a(reduce.o)
>      _MPIR_Reduce_intra in libmpich.a(reduce.o)
>      _MPIR_Reduce_intra in libmpich.a(reduce.o)
>      _MPIR_LAND in libmpich.a(opland.o)
>      _MPIR_LOR in libmpich.a(oplor.o)
>      _MPIR_MAXF in libmpich.a(opmax.o)
>      _MPIR_SUM in libmpich.a(opsum.o)
>      _MPIU_Strerror in libmpich.a(strerror.o)
>      _MPIU_Strerror in libmpich.a(strerror.o)
>      _MPIR_Reduce_scatter_block_intra in libmpich.a(red_scat_block.o)
>      _MPIR_Reduce_scatter_block_intra in libmpich.a(red_scat_block.o)
>      _MPIR_Reduce_scatter_block_intra in libmpich.a(red_scat_block.o)
> ld: symbol(s) not found
> 
> When I change the default Fortran compiler to GNU gfortran, the code compiles with no problems. I suspect this is related to an incompatibility between GCC and Intel Fortran Compiler. What do you think?
> 
> Thanks a lot!
> 
> _______________________________________________
> mpich-discuss mailing list
> mpich-discuss at mcs.anl.gov
> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
> 



More information about the mpich-discuss mailing list