<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
On 11/01/2012 9:18 PM, Ester Chiessi wrote:
<blockquote
cite="mid:1357D26CC5714767A44CAFA1902960D3@utente90e8f801"
type="cite">
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
<meta name="GENERATOR" content="MSHTML 8.00.6001.19170">
<style></style>
<div><font face="Arial" size="2">Hi,</font></div>
<div> </div>
<div><font face="Arial" size="2">running a md simulation with the
same tpr file ( using Gromacs 4.5.3 ) on three different
platforms ( A: <font face="Times New Roman" size="3"> </font><font
color="#000000" face="Times New Roman" size="3"><em>HP
xw8600 Workstation; B: PC with </em><font
face="Liberation Serif, serif"><font size="3"><font
color="#000000"><em>Intel CORE 2 DUO processor; C: </em>IBM
Power6 cluster ), I obtained the following results:</font></font></font></font></font></div>
<div> </div>
<div>on A and B platforms the run is fine,</div>
<div>on C the run crashes at the first step, returning LINCS
errors.</div>
<div> </div>
<div><font face="Arial" size="2">Did anyone observe this
behaviour? What does it mean?</font></div>
<div><font face="Arial" size="2">I wonder if my system is really
wrong or not.</font></div>
<div><font face="Arial" size="2">
<div> </div>
<div>I noted this difference in the corresponding md.log
files:</div>
<div>Platform A <span lang="IT">RMS relative
constraint deviation after constraining: 3.56e-06</span></div>
<div>Platform B <span lang="IT">RMS relative
constraint deviation after constraining: 3.48e-06</span></div>
<div>Platform C <span lang="IT">RMS relative
constraint deviation after constraining: 3.58e-06</span></div>
</font></div>
<div> </div>
<div><font face="Arial" size="2">Really many thanks for any
comment.</font></div>
<div> <br>
</div>
</blockquote>
<br>
See <a
href="http://www.gromacs.org/Documentation/Terminology/Reproducibility">http://www.gromacs.org/Documentation/Terminology/Reproducibility</a>
for general discussion. Probably A and B are getting lucky with the
numerical integration of a poorly conditioned system, and C is not.
For advice on fixing the problem, see <a
href="http://www.gromacs.org/Documentation/Terminology/Blowing_Up">http://www.gromacs.org/Documentation/Terminology/Blowing_Up</a><br>
<br>
Mark<br>
</body>
</html>