Newsgroups: comp.parallel.mpi
From: smith@epcc.ed.ac.uk (A Gordon Smith)
Subject: Re: real*4, T3D, MPI_REAL
Organization: Edinburgh Parallel Computing Centre
Date: Wed, 11 Oct 1995 13:21:23 GMT
Message-ID: <DGAD3o.Bzz.0.staffin.dcs.ed.ac.uk@dcs.ed.ac.uk>

In article <45f2cb$7bg@soenews.ucsd.edu>, impellus@ames.ucsd.edu (Tom Impelluso) writes:

> Hi, on the Cray T3D, I am using real*4 to ensure ieee 32 bit floats.
> Now I must decide if I pass things as MPI_REAL or MPI_DOUBLE_PRECISION.
> 
> But first, I must find out if MPI_REAL is 32 or 64 bit *IF*
> I have overridden the default with real*4
> 
> 
> Please let me know.
> 
> Thanks


The MPI Standard provides an optional basic datatype MPI_REAL4
to match the REAL*4 Fortran type -- see top p.19 of the 12/Jun/95
MPI Standard. This has been implemented in the CRI/EPCC MPI for
Cray T3D software.

-- 
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
 -=-=- A. Gordon Smith -=- Edinburgh Parallel Computing Centre -=-=-
 =-= Email <smith@epcc.ed.ac.uk> -=- Phone {+44 (0)131 650 6712} =-=
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=

