[mpich-discuss] Value of MPI_TAG_UB

[Bibrak Qamar]

Thanks Pavan !

Bibrak Qamar



On Thu, Mar 13, 2014 at 3:54 PM, Balaji, Pavan <balaji at anl.gov> wrote:

>
> Yes, the tag UB is correct.  MPICH reserves one of the tag bits to help
> with internal fault information propagation.
>
>   -- Pavan
>
> On Mar 13, 2014, at 1:30 AM, Bibrak Qamar <bibrakc at gmail.com> wrote:
>
> > I am trying to get the value of MPI_TAG_UB using the following code:
> >
> >   MPI_Aint  * maxTag;;
> >   int error, flag;
> >
> >   error = MPI_Comm_get_attr(MPI_COMM_WORLD, MPI_TAG_UB, &maxTag, &flag);
> >
> >   if(error != MPI_SUCCESS || !flag ){
> >   printf( " ERROR IN native mpitagub\n");
> >   }
> >
> >   printf( "mpitagub = %d\n", *maxTag);
> >
> >
> > For MPICH3 I get
> >
> > mpitagub = 1073741823
> >
> > I also did it for OpenMPI and I get
> >
> > mpitagub = 2147483647
> >
> > The value from OpenMPI makes sense since its 2^31 -1. But at the end the
> MPI_TAG_UB is MPI implementation dependent.
> >
> > So I was wondering the value that I got from MPICH3 (mpitagub =
> 1073741823) is this OK? or my code is wrong somewhere?
> >
> > Thanks
> > Bibrak
> >
> > _______________________________________________
> > discuss mailing list     discuss at mpich.org
> > To manage subscription options or unsubscribe:
> > https://lists.mpich.org/mailman/listinfo/discuss
>
> _______________________________________________
> discuss mailing list     discuss at mpich.org
> To manage subscription options or unsubscribe:
> https://lists.mpich.org/mailman/listinfo/discuss
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mpich.org/pipermail/discuss/attachments/20140313/f54dcafd/attachment.html>