[mpich-discuss] Profiling configuration file mpe_mpilog.conf

[Krishna Chaitanya]

Just out of curiosity, I was trying to measure the overhead due the MPE
component. For a message size of 0 bytes, averaged over  1000 attempts, I
noticed that the latency increases by about 150 micro s , when  compiled
with the -mpe=mpilog switch, Is this expected?


Thanks,
Krishna Chaitanya K


On Sat, May 3, 2008 at 12:55 PM, Krishna Chaitanya <kris.c1986 at gmail.com>
wrote:

> I think I figured it out. Its working now. Thanks.
>
> On Sat, May 3, 2008 at 2:42 AM, Anthony Chan <chan at mcs.anl.gov> wrote:
>
> >
> > ----- "Krishna Chaitanya" <kris.c1986 at gmail.com> wrote:
> >
> > > Hi,
> > > I have made a few changes in the MPI source code and added a few lines
> > > of code into the MPE module. To compile the cpi program in the
> > > examples directory, I have added the -mpe=mpilog switch in the
> > > makefile. It was working well until it started reporting this error :
> > >
> > > ../bin/mpicc -g -O2 -mpe=mpilog -L../lib -o cpi cpi.o -lm -lmpich
> > > -lpthread -lrt
> >
> > Assuming the mpicc that you invoked is in the mpich2 install directory,
> > you don't need "-L../lib" and "-lmpich ...." which are already done
> > in mpicc.
> >
> > > Profiling configuration file mpe_mpilog.conf not found in
> > > /home/mpich-install/etc
> >
> > Is there a mpe_mpilog.conf in /home/mpich-install/etc ?
> > If not, your installation of MPE2 is broken.
> >
> > > This is followed by a few undefined references to the function that I
> > > have added in the MPE module. I checked the messages displayed during
> > > make and I noticed that the libmpe.a library was created successfully
> > > at src/mpe2/lib. What could be the problem?
> >
> > Send us your configure, make and "make install" output as shown on
> > your screen.
> >
> > A.Chan
> >
> > > Thanks,
> > > Krishna Chaitanya,
> > >
> > >
> > >
> > > --
> > > In the middle of difficulty, lies opportunity
> >
> >
>
>
> --
> In the middle of difficulty, lies opportunity
>



-- 
In the middle of difficulty, lies opportunity
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/mpich-discuss/attachments/20080505/eeaacdf3/attachment.htm>