Newsgroups: comp.parallel.mpi
From: thakur@abacus.mcs.anl.gov (Rajeev Thakur)
Subject: derived datatypes
Organization: Argonne National Laboratory, Chicago, Illinois
Date: Tue, 12 May 1998 1:1:31 GMT
Message-ID: <89493489122009@abacus.mcs.anl.gov>


We recently wrote the following paper, which I thought may be of
interest to some.

Rajeev

    
     A Case for Using MPI's Derived Datatypes to Improve I/O Performance
     -------------------------------------------------------------------

               Rajeev Thakur, William Gropp, and Ewing Lusk
                       Argonne National Laboratory

                   Preprint ANL/MCS-P717-0598, May 1998

        
                             Abstract 
                             --------
             
MPI-IO, the I/O part of the MPI-2 standard, is a promising new
interface for parallel I/O. A key feature of MPI-IO is that it allows
users to access several noncontiguous pieces of data from a file with
a single I/O function call by defining file views with derived
datatypes. We explain how critical this feature is for high
performance, why users must create and use derived datatypes whenever
possible, and how it enables implementations to perform optimizations.
In particular, we describe two optimizations our MPI-IO
implementation, ROMIO, performs: data sieving and collective I/O. We
present performance results on five different parallel machines: HP
Exemplar, IBM SP, Intel Paragon, NEC SX-4, and SGI Origin2000.


Paper available at 
http://www.mcs.anl.gov/home/thakur/dtype.ps.gz   (postscript version)
or 
http://www.mcs.anl.gov/home/thakur/dtype    (html version)