From: Luc.Vereecken@chem.kuleuven.ac.be (Luc Vereecken)
Newsgroups: comp.parallel.mpi
Subject: Re: Pentium related problem
Date: Sat, 12 Sep 1998 11:53:23 GMT
Organization: KULeuvenNet
Message-Id: <35f925a0.702274345@news.kuleuven.ac.be>
References: <35F84687.69A9365@scl.ameslab.gov>
Mime-Version: 1.0
Content-Type: text/plain; charset=iso-8859-1
Content-Transfer-Encoding: 8bit
Cache-Post-Path: marvin!unknown@pcjp1.chem.kuleuven.ac.be


On Thu, 10 Sep 1998 16:37:12 -0500, Vasiliu Bogdan
<bogdan@scl.ameslab.gov> wrote:

>The strange thing is that the unpacking is one order
>of magnitude slower than the packing.  I do not understand
>why the unpacking process (which is the inverse of the packing)
>is 10 times slower than the packing.
>I wrote a small C test program (without any communication calls) and
>I did some tests on SGI, HP, and Pentium computers. Only the Pentiums
>had this problem  (this process of packing and unpacking data occurs
>in any parallel FDM algorithm).
>Does anyone know the answer to this problem?
>

Maybe this is just an alignment problem. Maybe the compiler you use on
the pentiums doesn't optimise enough, such that one/all of the buffers
are not on the proper boundary. You use an explicit malloc for
packing, but don't for the unpacking, such that maybe the temp array
has a different alignment. This should be easy to check by examining
the offset-part of a pointer to the structure. 

Luc Vereecken