[mvapich-discuss] Segmentation fault when compiling with Intel 15, mvapich2 2.1, and openmp

Thu Sep 10 17:23:14 EDT 2015

Hi Craig.

There has been extensive discussion about this issue between TACC and
Intel. To the best of our
knowledge, the issue is been localized to the Intel compiler.  This seems
to affect versions 15.0.2 and up.

On Thu, Sep 10, 2015 at 4:07 PM, Craig Tierney - NOAA Affiliate <
craig.tierney at noaa.gov> wrote:

> --===============3875389897677332204==
> Content-Type: multipart/alternative;
> boundary="047d7bdca5c272ca21051f6a262e"
>
> --047d7bdca5c272ca21051f6a262e
> Content-Type: text/plain; charset="UTF-8"
>
> Hello,
>
> I am having a problem building and running codes with Intel 15.X and
> Mvapich2 2.1 that use openmp.  The following code shows the problem:
>
> program test
>     use mpi
>     integer :: ierr, iprov
>     real(8) :: a
>     call mpi_init_thread(MPI_THREAD_FUNNELED,iprov,ierr)
>     call random_number(a)
>     write(*,*)"hello"
>     call mpi_finalize(ierr)
> end program test
>
> If this is compiled as:
>
> # mpif90 -openmp -O3 test.f90 -o test
>
> And run, the following error message is generated.
>
> # mpiexec.hydra -np 2 ./testmpifort
>
>
> ===================================================================================
> =   BAD TERMINATION OF ONE OF YOUR APPLICATION PROCESSES
> =   PID 17905 RUNNING AT fe8
> =   EXIT CODE: 11
> =   CLEANING UP REMAINING PROCESSES
> =   YOU CAN IGNORE THE BELOW CLEANUP MESSAGES
>
> ===================================================================================
> YOUR APPLICATION TERMINATED WITH THE EXIT STRING: Segmentation fault
> (signal 11)
> This typically refers to a problem with your application.
> Please see the FAQ page for debugging suggestions
>
> When you get the stack trace, you see:
>
> (gdb) bt
> #0  0x000000000040b0c4 in init_resource ()
> #1  0x000000000040b02a in reentrancy_init ()
> #2  0x000000000040af48 in for__reentrancy_init ()
> #3  0x00002aaaab329115 in for_rtl_init_ () from
> /home/admin/software/apps/mvapich2/2.1-intel/lib/libmpi.so.12
> #4  0x0000000000403249 in main ()
>
> Intel believes this is a mvapich2 issue, not a compiler issue.
>
>
> https://software.intel.com/en-us/forums/intel-fortran-compiler-for-linux-and-mac-os-x/topic/540673
>
> Thanks,
> Craig
>
> --047d7bdca5c272ca21051f6a262e
> Content-Type: text/html; charset="UTF-8"
> Content-Transfer-Encoding: quoted-printable
>
> <div dir=3D"ltr">Hello,<div><br></div><div>I am having a problem building
> a=
> nd running codes with Intel 15.X and Mvapich2 2.1 that use openmp.=C2=A0
> Th=
> e following code shows the problem:</div><div><br></div><div><div>program
> t=
> est</div><div>=C2=A0 =C2=A0 use mpi</div><div>=C2=A0 =C2=A0 integer ::
> ierr=
> , iprov</div><div>=C2=A0 =C2=A0 real(8) :: a</div><div>=C2=A0 =C2=A0 call
> m=
> pi_init_thread(MPI_THREAD_FUNNELED,iprov,ierr)</div><div>=C2=A0 =C2=A0
> call=
>  random_number(a)</div><div>=C2=A0 =C2=A0
> write(*,*)"hello"</div>=
> <div>=C2=A0 =C2=A0 call mpi_finalize(ierr)</div><div>end program
> test</div>=
> </div><div><br></div><div>If this is compiled
> as:</div><div><br></div><div>=
> # mpif90 -openmp -O3 test.f90 -o test</div><div><br></div><div>And run,
> the=
>  following error message is generated.</div><div><br></div><div>#
> mpiexec.h=
> ydra -np 2 ./testmpifort</div><div><br></div><div>=3D=3D=3D=3D=3D=3D=3D=3D=
>
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
>
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
>
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
> </div><div>=3D =C2=A0 BAD TERMINATION OF ONE OF YOUR APPLICATION
> PROCESSES<=
> /div><div>=3D =C2=A0 PID 17905 RUNNING AT fe8</div><div>=3D =C2=A0 EXIT
> COD=
> E: 11</div><div>=3D =C2=A0 CLEANING UP REMAINING PROCESSES</div><div>=3D =
> =C2=A0 YOU CAN IGNORE THE BELOW CLEANUP MESSAGES</div><div>=3D=3D=3D=3D=3D=
>
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
>
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
>
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
> =3D=3D=3D</div><div>YOUR APPLICATION TERMINATED WITH THE EXIT STRING:
> Segme=
> ntation fault (signal 11)</div><div>This typically refers to a problem
> with=
>  your application.</div><div>Please see the FAQ page for debugging
> suggesti=
> ons</div><div><br></div><div>When you get the stack trace, you
> see:</div><d=
> iv><br></div><div><div>(gdb) bt</div><div>#0 =C2=A00x000000000040b0c4 in
> in=
> it_resource ()</div><div>#1 =C2=A00x000000000040b02a in reentrancy_init
> ()<=
> /div><div>#2 =C2=A00x000000000040af48 in for__reentrancy_init
> ()</div><div>=
> #3 =C2=A00x00002aaaab329115 in for_rtl_init_ () from
> /home/admin/software/a=
> pps/mvapich2/2.1-intel/lib/libmpi.so.12</div><div>#4
> =C2=A00x00000000004032=
> 49 in main ()</div></div><div><br></div><div>Intel believes this is a
> mvapi=
> ch2 issue, not a compiler issue.</div><div><br></div><div><a
> href=3D"https:=
> //
> software.intel.com/en-us/forums/intel-fortran-compiler-for-linux-and-mac-=
> os-x/topic/540673">
> https://software.intel.com/en-us/forums/intel-fortran-co=
> mpiler-for-linux-and-mac-os-x/topic/540673
> </a><br></div><div><br></div><div=
> >Thanks,</div><div>Craig</div></div>
>
> --047d7bdca5c272ca21051f6a262e--
>
> --===============3875389897677332204==
> Content-Type: text/plain; charset="us-ascii"
> MIME-Version: 1.0
> Content-Transfer-Encoding: 7bit
> Content-Disposition: inline
>
> _______________________________________________
> mvapich-discuss mailing list
> mvapich-discuss at cse.ohio-state.edu
> http://mailman.cse.ohio-state.edu/mailman/listinfo/mvapich-discuss
>
> --===============3875389897677332204==--
>

-- 
Jonathan Perkins
http://www.cse.ohio-state.edu/~perkinjo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.cse.ohio-state.edu/pipermail/mvapich-discuss/attachments/20150910/35eb5071/attachment-0001.html>