Newsgroups: comp.parallel.mpi
From: Christopher Ditchman <cjd@yosemite.jpl.nasa.gov>
Reply-To: cjd@yosemite.jpl.nasa.gov
Subject: Re: Possible Memory Leak in MPI update
Organization: Jet Propulsion Laboratory
Date: Thu, 11 Jul 1996 09:24:34 +0000
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Message-ID: <31E4C84D.4114@yosemite.jpl.nasa.gov>

Hello,

Thanks to everyone for their responses.  To clarify,
I am using CRI/EPCC MPI version 1.4a  

I am in contact with local T3D consultants and also
the ones from EPCC.  I am also curious to see if 
someone might have an idea about a work around for my problem.

I am making use of a software package called Parallel
Genetic Algorithm package put out by Argnonne National
Lab.  The package continually creates and frees a derived
datatype which points to a specific element in an array of
structures.  This derived type consists of pointers to 
four fields in the structure for the particular element for
which the datatype was created. The package is written in C.
From what I gather, there is little support currently for the
package and will not likely have much in the future. 

C is not my strong language.   I would like to see if I could
modify the package and get around having to continually create
and free the datatype.  It has been suggested that I make use
of static variables to match MPI datatypes to the data structures
in my program.  Could someone give me an example of how to 
go about this?  

Thanks,

Chris

-----
Christopher J. Ditchman			Tracking Systems and Applications Group
cjd@yosemite.jpl.nasa.gov		Jet Propulsion Laboratory
							4800 Oak Grove Dr.
							Pasadena, CA 91109 MS 238-600

