[mpich-discuss] -mpe=mpitrace not producing any output, and Valgrind outputs on r3717

[Anthony Chan]

----- "James S Perrin" <james.s.perrin at manchester.ac.uk> wrote:

> Hi,
> 
> 	I too am having problems with -mpe=mpitrace, ironically I was trying
> to use it to analyse the speed issue with nemesis. All I get in the way
> of output is :
> 
> [0] Ending MPI_Finalize
> [1] Ending MPI_Finalize
> [3] Ending MPI_Finalize
> [2] Ending MPI_Finalize
> [4] Ending MPI_Finalize
>
> I can't downsize my code unfortunately and I get start and end
> messages for MPI calls when I compile cpi.

Does your code use any MPI-2 features, like connect/accept or
comm_spawn or even put/get ?  Did you try -mpe=mpilog ?

> However if this option is equivalent the the MPE -mpitrace option
> 
> http://www-unix.mcs.anl.gov/mpi/mpich1/docs/mpeman/node7.htm#Node8
> 
> should it also be printing out some accumulated times as well?

I look up some really old versions of trc_wrappers.c in MPICH1 and I don't
see it has any statistics collection code in it.  My guess is that the
MPICH1 documentation was never updated.  BTW, if you need simple statistics,
you could use FPMPI2 at http://www.mcs.anl.gov/fpmpi 
(the latest version about 1 year old)

A.Chan

> Regards
> James
> 
> François PELLEGRINI wrote:
> > Hello and happy new year to all,
> > 
> > Several topics are addressed in this e-mail.
> > 
> > First, I have trouble using the "mpitrace" feature of MPE.
> > I compile all of my source code with "mpicc -mpe=mpitrace",
> > and link the objects with "-ltmpe", but absolutely no output
> > is produced when I run the compiled program. This both happens
> > with the official 1.0.8 and r3717 packages. What did I do
> > wrong ?
> 
> -- 
> ------------------------------------------------------------------------
>    James S. Perrin
>    Visualization
> 
>    Research Computing Services
>    Devonshire House, University Precinct
>    The University of Manchester
>    Oxford Road, Manchester, M13 9PL
> 
>    t: +44 (0) 161 275 6945
>    e: james.perrin at manchester.ac.uk
>    w: www.manchester.ac.uk/researchcomputing
> ------------------------------------------------------------------------
>   "The test of intellect is the refusal to belabour the obvious"
>   - Alfred Bester
> ------------------------------------------------------------------------