<html>
<head>
<style><!--
.hmmessage P
{
margin:0px;
padding:0px
}
body.hmmessage
{
font-size: 10pt;
font-family:Verdana
}
--></style>
</head>
<body class='hmmessage'>
<br><br>> Date: Wed, 7 Oct 2009 15:51:00 +0200<br>> From: jochen@xray.bmc.uu.se<br>> To: gmx-users@gromacs.org<br>> Subject: Re: [gmx-users] vsites and lincs-order<br>> <br>> All right, thanks!<br>> <br>> Since you mention energy conservation, maybe it would be worth adding <br>> notes or a warnings into pdb2gmx if the time step is large. First, if <br>> lincs-order is 4 (instead of 6) and second (more important), if nstlist <br>> is not reduced with increasing dt. I strongly feel that most people who <br>> use vsites forget to reduce nstlist accordingly.<br>> <br><br>grompp you mean, I guess?<br><br>> I have played a bit with timestep and nstlist and checked the energy <br>> drift - apparently neighbor searching should be done at least every <br>> 20fs. With less frequent neighbor searching the energy drift increases <br>> drastically (with PME and LJ-cutoff, what people typically use). Now <br>> imagine a time step of 5fs, with the standard nstlist=10, yields <br>> neighbor searching only every 50fs. In my test system that increased the <br>> energy drift by a factor of 10-20 compared to the case where neighbor <br>> searching was done every 20fs.<br><br>We can not do such things.<br>Gromacs is not a atomistic, protein specific simulations package.<br>For other systems, masses, etc. such settings could be fine.<br>Moreover, grompp can not easily check for this specific type of vsite use.<br>TIP4p, for instance, also contains a vsite.<br><br>It has happened that someone put in similar checks in grompp, but this can be<br>EXTREMELY annoying when you use different systems.<br><br>Berk<br><br>> <br>> Jochen<br>> <br>> <br>> Berk Hess wrote:<br>> > My P-Lincs paper http://dx.doi.org/10.1021/ct700200b shows that<br>> > with order 6 and a time step of 4 fs you get roughly the same <br>> > constraint accuracy<br>> > and energy conservation as without v-sites and a 2 fs time step.<br>> > With order 4 and a 4 fs time step the energy drift is 2.2 times higher <br>> > than with order 6<br>> > (not mentioned in the paper). This is not a very large difference.<br>> ><br>> > I recall that some time ago I decided to change the default lincs <br>> > order to 6<br>> > (I have not done this yet), but unfortunately now I don't recall what <br>> > issue<br>> > made me decide that the current default accuracy is not high enough.<br>> ><br>> > Berk<br>> ><br>> > > Date: Wed, 7 Oct 2009 13:16:52 +0200<br>> > > From: jochen@xray.bmc.uu.se<br>> > > To: gmx-users@gromacs.org<br>> > > Subject: [gmx-users] vsites and lincs-order<br>> > ><br>> > > Hi,<br>> > ><br>> > > the manual suggests lincs-order = 6 when using large time steps (4-5 <br>> > fs,<br>> > > with vsites). Has anyone experience how severe that issue is. Has <br>> > anyone<br>> > > observed artefacts with lincs-order=4 and large time steps?<br>> > ><br>> > > Thanks a lot,<br>> > ><br>> > > Jochen<br>> > ><br>> > > --<br>> > > ---------------------------------------------------<br>> > > Dr. Jochen Hub<br>> > > Molecular Biophysics group<br>> > > Dept. of Cell & Molecular Biology<br>> > > Uppsala University. Box 596, 75124 Uppsala, Sweden.<br>> > > Phone: +46-18-4714451 Fax: +46-18-511755<br>> > > ---------------------------------------------------<br>> > ><br>> > > _______________________________________________<br>> > > gmx-users mailing list gmx-users@gromacs.org<br>> > > http://lists.gromacs.org/mailman/listinfo/gmx-users<br>> > > Please search the archive at http://www.gromacs.org/search before <br>> > 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>> > What can you do with the new Windows Live? Find out <br>> > <http://www.microsoft.com/windows/windowslive/default.aspx><br>> > ------------------------------------------------------------------------<br>> ><br>> > _______________________________________________<br>> > gmx-users mailing list gmx-users@gromacs.org<br>> > http://lists.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>> -- <br>> ---------------------------------------------------<br>> Dr. Jochen Hub<br>> Molecular Biophysics group<br>> Dept. of Cell & Molecular Biology<br>> Uppsala University. Box 596, 75124 Uppsala, Sweden.<br>> Phone: +46-18-4714451 Fax: +46-18-511755<br>> ---------------------------------------------------<br>> <br>> _______________________________________________<br>> gmx-users mailing list gmx-users@gromacs.org<br>> http://lists.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' target='_new'>Find out</a></body>
</html>