Hello everyone,
I am a lam mpi user for approximately 2 years and i
have been experimenting over many cluster types.
During these period of time, i found almost no problem
with lam, especially in some MPI-2 characteristics, in
which many other mpi distributions presented lots of
problems (i didn't have the chance to use open-mpi
yet). I would also like to congratulate those who
answer to the virtually infinite amount of problems in
the lam list and I hope someone will help me, as well
:-). My problems are the following:
i) We have an MPMD application with 2 executables,
let's say A and B. When we want to run the
application, we use mpirun with an appschema file, in
order to put all processes under the same
MPI_COMM_WORLD. In this file we have:
-np <number-of-procs-a> A <some arguments>
-np <number-of-procs-b> B <some other arguments>
The question is this: how can we attach a debugger to
those processes (to process 0 or even to all of them)?
I have tried many methods, based on the debugging info
in the lam documentation, but nothing seemed to work.
ii) The reason that i want to use the debugger in this
particular scenario (MPMD application), is that our
MPMD application runs without any errors on a single
machine, but fails to run on two (or more) machines.
The error is located when the MPI_Comm_split function
is called. However, we have tested this function
thoroughly on small test programs and on both single
and multiple machine execution and everything worked
just fine. So, most probably the problem occurs
because of some memory leak. Is there any possibility
that the nature of this problem is not connected to
memory leaks?
Thank you in advance,
Paul Polydoras,
University of Athens, Department of Informatics and
Telecommunications, Parallel Lab.
___________________________________________________________
×ñçóéìïðïéåßôå Yahoo!;
ÂáñåèÞêáôå ôá åíï÷ëçôéêÜ ìçíýìáôá (spam); Ôï Yahoo! Mail
äéáèÝôåé ôçí êáëýôåñç äõíáôÞ ðñïóôáóßá êáôÜ ôùí åíï÷ëçôéêþí
ìçíõìÜôùí http://login.yahoo.com/config/mail?.intl=gr
|
