Newsgroups: comp.parallel.mpi
From: salo@mrjones.engr.sgi.com (Eric Salo)
Subject: Re: Performance Of MPI_WTIME()
Organization: Silicon Graphics, Inc.  Mountain View, CA
Date: 21 Oct 1996 20:14:09 GMT
Message-ID: <54glih$a32@murrow.corp.sgi.com>

> According to "MPI: The Complete Reference" by M. Snir et al. MPI_WTIME()
> "carries no unnecessary baggage." Does this mean that even if I use it a
> lot the overall performance will not be affected much? Is this also true
> in practice?

I believe that "no unnecessary baggage" refers to the fact that, at least
in principle, it should be possible to implement this call with minimal
overhead. This is probably more true in practice for the vendor
implementations than for the portable implementations; in our case, it's
a single uncached load from a hardware timer followed by a
floating-point multiply.

Eric Salo         Silicon Graphics Inc.             "Do you know what the
(415)933-2998     2011 N. Shoreline Blvd, 8U-802     last Xon said, just
salo@sgi.com      Mountain View, CA   94043-1389     before he died?"

