<html>
<head>
<style>
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 10pt;
font-family:Verdana
}
</style>
</head>
<body class='hmmessage'>
Hi,<br><br>I added a section on Exact vs binary identical continuation on the extending simulations wiki page.<br><br>Berk<br><br><hr id="stopSpelling">From: gmx3@hotmail.com<br>To: gmx-users@gromacs.org<br>Subject: RE: [gmx-users] continuing a MD simulation on different architectures<br>Date: Thu, 2 Apr 2009 15:20:03 +0200<br><br>
<style>
.ExternalClass .EC_hmmessage P
{padding:0px;}
.ExternalClass body.EC_hmmessage
{font-size:10pt;font-family:Verdana;}
</style>
Hi,<br><br>If you had a computer with unlimited precision, or if you integrated the time discretized<br>equations of motion by hand, exact continuation would lead to identical results.<br>But since computer have limited precision (single or double precision) and MD is chaotic,<br>trajectories will diverge very rapidly if one bit (the least significant one) is rounded differently.<br>Such trajectories will all be equally valid, but very different.<br>In most cases you don't care that a trajectory does not run exactly identical when you stop<br>it and continue it, as opposed to keeping it running in a single run.<br>The most common case where you do care is when you have a crash at e.g. step 567894 and<br>you want to reproduce it to track down if it is a problem with your system or a bug.<br><br>A discussion like this should be added to the wiki.<br>I just noticed that currently there are two pages (doing restarts and extending simjulations)<br>that should be merged.<br><br>Berk<br><br>> Date: Thu, 2 Apr 2009 15:04:13 +0200<br>> From: spitaleri.andrea@hsr.it<br>> To: gmx-users@gromacs.org<br>> Subject: Re: [gmx-users] continuing a MD simulation on different architectures<br>> <br>> Hi,<br>> thanks for the answer. One thing: what do you mean for "but not binary identical"?<br>> >From the manual mdrun says:<br>> "The only disadvantage of dynamic load balancing is that runs are no longer binary reproducible, but<br>> in most cases this is not important"<br>> <br>> this sentence is also not very clear to me.<br>> <br>> thanks in advance<br>> <br>> Regards<br>> <br>> andrea<br>> <br>> <br>> Berk Hess wrote:<br>> > Hi,<br>> > <br>> > As mdrun will tell you in stderr and md.log when you try continuing,<br>> > you will have exact continuation, but not binary identical (which you would<br>> > not have anyhow if you use dynamic load balancing).<br>> > Exact continuation mean perfect physical and chemical reliability.<br>> > <br>> > Berk<br>> > <br>> > <br>> >> Date: Thu, 2 Apr 2009 13:28:21 +0200<br>> >> From: spitaleri.andrea@hsr.it<br>> >> To: gmx-users@gromacs.org<br>> >> Subject: [gmx-users] continuing a MD simulation on different architectures<br>> >><br>> >> Dear all,<br>> >> is there any problem to continue (extend) a simulation using a different architecture hardware?<br>> >> I have a set of simulations coming from xt4 and xt5 architecture (eepc and csc clusters) and I would<br>> >> like to extend the simulation step on my cluster (intel) using the previous cpt file and generating<br>> >> the new tpr on my cluster. I did it and it runs, however I was wondering about the physical and<br>> >> chemical reliability.<br>> >><br>> >> thanks in advance<br>> >><br>> >> Regards<br>> >><br>> >> andrea<br>> >><br>> >> --<br>> >> -------------------------------<br>> >> Andrea Spitaleri PhD<br>> >> Dulbecco Telethon Institute<br>> >> c/o DIBIT Scientific Institute<br>> >> Biomolecular NMR, 1B4<br>> >> Via Olgettina 58<br>> >> 20132 Milano (Italy)<br>> >> http://biomolecularnmr.ihsr.dom/<br>> >> Tel: 0039-0226434348/5622/3497/4922<br>> >> Fax: 0039-0226434153<br>> >> -------------------------------<br>> >><br>> >> -----------------------------------------------------------------<br>> >><br>> >> La tua mano puo' lasciare un segno importante.<br>> >> Dona il tuo 5 per mille al San Raffaele di Milano.<br>> >><br>> >> E' SEMPLICE E NON COSTA NULLA.<br>> >> Basta indicare nell'apposito riquadro della dichiarazione dei redditi "Finanziamento della ricerca sanitaria"<br>> >> il codice fiscale della Fondazione Centro S. Raffaele del Monte Tabor:<br>> >> 03 06 42 80 153 e ricordarsi di firmare.<br>> >> Per saperne di piu': 5permille@hsr.it o vai sul sito http://www.5xmille.org.<br>> >> _______________________________________________<br>> >> gmx-users mailing list gmx-users@gromacs.org<br>> >> http://www.gromacs.org/mailman/listinfo/gmx-users<br>> >> Please search the archive at http://www.gromacs.org/search before posting!<br>> >> Please don't post (un)subscribe requests to the list. Use the<br>> >> www interface or send it to gmx-users-request@gromacs.org.<br>> >> Can't post? Read http://www.gromacs.org/mailing_lists/users.php<br>> > <br>> > ________________________________<br>> > Express yourself instantly with MSN Messenger! MSN Messenger<http://clk.atdmt.com/AVE/go/onm00200471ave/direct/01/><br>> > <br>> <br>> -- <br>> -------------------------------<br>> Andrea Spitaleri PhD<br>> Dulbecco Telethon Institute<br>> c/o DIBIT Scientific Institute<br>> Biomolecular NMR, 1B4<br>> Via Olgettina 58<br>> 20132 Milano (Italy)<br>> http://biomolecularnmr.ihsr.dom/<br>> Tel: 0039-0226434348/5622/3497/4922<br>> Fax: 0039-0226434153<br>> -------------------------------<br>> <br>> -----------------------------------------------------------------<br>> <br>> La tua mano puo' lasciare un segno importante.<br>> Dona il tuo 5 per mille al San Raffaele di Milano.<br>> <br>> E' SEMPLICE E NON COSTA NULLA.<br>> Basta indicare nell'apposito riquadro della dichiarazione dei redditi "Finanziamento della ricerca sanitaria"<br>> il codice fiscale della Fondazione Centro S. Raffaele del Monte Tabor:<br>> 03 06 42 80 153 e ricordarsi di firmare.<br>> Per saperne di piu': 5permille@hsr.it o vai sul sito http://www.5xmille.org.<br>> _______________________________________________<br>> gmx-users mailing list gmx-users@gromacs.org<br>> http://www.gromacs.org/mailman/listinfo/gmx-users<br>> Please search the archive at http://www.gromacs.org/search before posting!<br>> Please don't post (un)subscribe requests to the list. Use the <br>> www interface or send it to gmx-users-request@gromacs.org.<br>> Can't post? Read http://www.gromacs.org/mailing_lists/users.php<br><br><hr>What can you do with the new Windows Live? <a href="http://www.microsoft.com/windows/windowslive/default.aspx">Find out</a><br /><hr />Express yourself instantly with MSN Messenger! <a href='http://clk.atdmt.com/AVE/go/onm00200471ave/direct/01/' target='_new'>MSN Messenger</a></body>
</html>