I don&#39;t really care what others say, but to prove that this has any performance value you should do the following:<div><br></div><div><br></div><div>Compare your most &quot;parallel&quot; algorithm with the performance of a corresponding well-written MPI application using openmpi&#39;s shared memory transport. If there is a difference, then your system has some value.</div>
<div><br></div><div>Of course openmpi&#39;s shared memory transport is terribly buggy, but it should give a baseline acceptable performance.</div><div><br></div><div>If there is no comparison, we have no idea.</div><div><br>
</div><div>Best,<br><br clear="all"><br>-- <br>Eray Ozkural, PhD candidate.  Comp. Sci. Dept., Bilkent University, Ankara<br><a href="http://groups.yahoo.com/group/ai-philosophy">http://groups.yahoo.com/group/ai-philosophy</a><br>
<a href="http://myspace.com/arizanesil">http://myspace.com/arizanesil</a> <a href="http://myspace.com/malfunct">http://myspace.com/malfunct</a><br><br>
</div>