<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=us-ascii">
</head>
<body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; font-family: Calibri, sans-serif; font-size: 14px; color: rgb(0, 0, 0);">
<div><br>
</div>
<blockquote id="MAC_OUTLOOK_ATTRIBUTION_BLOCKQUOTE" style="BORDER-LEFT: #b5c4df 5 solid; PADDING:0 0 0 5; MARGIN:0 0 0 5;">
<div>Yes: periodic molecules </div>
</blockquote>
<div><br>
</div>
<div>I think this is mostly solved by coupled constraint system COM constraints (I may be wrong), unless the entire periodic molecule is constrained, which seems like an unlikely case to be running MD on.</div>
<div><br>
</div>
<div>> and COM (pull) constraints.</div>
<div><br>
</div>
<div>To avoid communication, scaling would need to preserve the distance between these atoms. Although the distance between the two is known by all nodes (it's a constant), the direction between the constraints is not known, so one wouldn't know where the
center of that constrained system was -- so I guess communication would be required in this case. </div>
<div><br>
</div>
<div>But maybe I'm totally off, and if a constraint system is split between multiple nodes nodes, there is no easy way to compute the COM of that constraint w/o communication, and one wouldhave to communicate anyway.</div>
<div><br>
</div>
<div>Best,</div>
<div>~~~~~~~~~~~~</div>
<div>Michael Shirts</div>
<div>Assistant Professor</div>
<div>Department of Chemical Engineering</div>
<div>University of Virginia</div>
<div><a href="mailto:michael.shirts@virginia.edu">michael.shirts@virginia.edu</a></div>
<div>(434)-243-1821</div>
<div><br>
</div>
</body>
</html>