<html>
<head>
<style>
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 10pt;
font-family:Verdana
}
</style>
</head>
<body class='hmmessage'>
<div><br></div><div>Don't forget matching. The model depends on a relation between send and receive. This is the fundamental reason for potential difference in overlap. If you talk implementation, which we technically shouldn't for this argument, eventually the fact that a matching receiver is required for a send does impact over multiple sends. <br></div><div><br></div><div>the one-sided model</div><div><br></div><div>a->b </div><div><br></div><div>is independent of b. </div><div><br></div><div>the two sided model</div><div><br></div><div>a<->b</div><div><br></div><div>because of its dependance implies validity -- you can hide the cost of validity but can't eliminate it. </div><div><br></div><div>Vinod.</div><div><br>> From: keith.d.underwood@intel.com<br>> To: mpi-forum@lists.mpi-forum.org<br>> Date: Mon, 27 Apr 2009 06:48:57 -0600<br>> Subject: Re: [Mpi-forum] MPI One-Sided Communication<br>> <br>> <br>>>On the Earth Simulator, there are/were several application codes which are<br>>>using one-sided communication (instead of 2-sided). They used one-sided<br>>>communication especially to overlap communication and computation.<br>>>When I remember correctly, at least one of this applications won a Gordon<br>>>Bell Award of SC.<br>> <br>> The ambiguity of the progress rule notwithstanding, there is no particular reason that one-sided should give you better overlap than two-sided. If this is the reason that people use one-sided, maybe we should revisit the progress rule ;-)<br>> <br>> Keith<br>> <br>> _______________________________________________<br>> mpi-forum mailing list<br>> mpi-forum@lists.mpi-forum.org<br>> http://lists.mpi-forum.org/mailman/listinfo.cgi/mpi-forum<br></div></body>
</html>