Open MPI logo

Open MPI User's Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Open MPI User's mailing list

Subject: Re: [OMPI users] MPI_TAG_UB printing zero with Intel Compiler
From: Jeffrey Squyres (jsquyres_at_[hidden])
Date: 2012-03-14 08:22:34


George is correct. Try this:

[5:21] svbu-mpi:~/mpi % cat tag-ub.f90
program mpitag
      use mpi
      implicit none
      integer :: err
      integer (KIND=MPI_ADDRESS_KIND) :: my_tag_ub
      logical flag

      call MPI_INIT(err)
      call MPI_COMM_GET_ATTR(MPI_COMM_WORLD, MPI_TAG_UB, my_tag_ub, flag, err)
      if (flag .eq. .true.) then
         print *, "Got tag ub:", my_tag_ub
      else
         print *, "Couldn't find tag ub!"
      endif
      call MPI_FINALIZE(err)
end program mpitag
[5:21] svbu-mpi:~/mpi % mpif90 tag-ub.f90
[5:22] svbu-mpi:~/mpi % ./a.out
 Got tag ub: 2147483647
[5:22] svbu-mpi:~/mpi %

On Mar 13, 2012, at 7:50 PM, George Bosilca wrote:

> MPI_TAG_UB is not a constant, it is a predefined attribute. As such it should be accessed using the attribute accessors (MPI_COMM_GET_ATTR page 229 in the MPI 2.2 standard).
>
> george.
>
> On Mar 13, 2012, at 13:51 , Timothy Stitt wrote:
>
>> Hi Jeff,
>>
>> I went through the procedure of compiling and running, then copied the procedure verbatim from the command line (see below).
>>
>> ********
>> [tstitt_at_memtfe] /pscratch/tstitt > more mpitag.f90
>> program mpitag
>>
>> use mpi
>>
>> implicit none
>>
>> integer :: err
>>
>> call MPI_INIT(err)
>>
>> print *,MPI_TAG_UB
>>
>> call MPI_FINALIZE(err)
>>
>> end program mpitag
>> [tstitt_at_memtfe] /pscratch/tstitt > mpif90 --version
>> ifort (IFORT) 12.0.2 20110112
>> Copyright (C) 1985-2011 Intel Corporation. All rights reserved.
>>
>> [tstitt_at_memtfe] /pscratch/tstitt > mpiexec --version
>> mpiexec (OpenRTE) 1.4.3
>>
>> Report bugs to http://www.open-mpi.org/community/help/
>> [tstitt_at_memtfe] /pscratch/tstitt > mpif90 -o mpitag mpitag.f90
>> [tstitt_at_memtfe] /pscratch/tstitt > mpiexec -n 2 ./mpitag
>> 0
>> 0
>> **********
>>
>> Let me know if you need any further information.
>>
>> Thanks,
>>
>> Tim.
>>
>> On Mar 13, 2012, at 1:39 PM, Jeffrey Squyres wrote:
>>
>>> Tim --
>>>
>>> I am unable to replicate this problem with a 1.4 build with icc.
>>>
>>> Can you share your test code?
>>>
>>>
>>> On Mar 10, 2012, at 7:30 PM, Timothy Stitt wrote:
>>>
>>>> Hi all,
>>>>
>>>> I was experimenting with MPI_TAG_UB in my code recently and found that its value is set to 0 in my v1.4.3 and v1.4.5 builds (these are the only builds I have available to test).
>>>>
>>>> Oddly it only happens with my builds compiled with the Intel compiler. The PGI builds do produce a non-zero value. Has anyone else experienced this?
>>>>
>>>> Regards,
>>>>
>>>> Tim.
>>>> _______________________________________________
>>>> users mailing list
>>>> users_at_[hidden]
>>>> http://www.open-mpi.org/mailman/listinfo.cgi/users
>>>
>>>
>>> --
>>> Jeff Squyres
>>> jsquyres_at_[hidden]
>>> For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/
>>>
>>>
>>> _______________________________________________
>>> users mailing list
>>> users_at_[hidden]
>>> http://www.open-mpi.org/mailman/listinfo.cgi/users
>>
>> Tim Stitt PhD (User Support Manager).
>> Center for Research Computing | University of Notre Dame |
>> P.O. Box 539, Notre Dame, IN 46556 | Phone: 574-631-5287 | Email: tstitt_at_[hidden]
>>
>> _______________________________________________
>> users mailing list
>> users_at_[hidden]
>> http://www.open-mpi.org/mailman/listinfo.cgi/users
>
> _______________________________________________
> users mailing list
> users_at_[hidden]
> http://www.open-mpi.org/mailman/listinfo.cgi/users

-- 
Jeff Squyres
jsquyres_at_[hidden]
For corporate legal information go to: http://www.cisco.com/web/about/doing_business/legal/cri/