<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=Windows-1252">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<style type="text/css" style="display:none;"><!-- P {margin-top:0;margin-bottom:0;} --></style>
<div id="divtagdefaultwrapper" style="font-size:12pt;color:#000000;font-family:Calibri,Arial,Helvetica,sans-serif;" dir="ltr">
<p>I am referring to the <font size="3" face="Calibri,Arial,Helvetica,sans-serif" color="black">
<span style="font-size:12pt;" id="divtagdefaultwrapper">original Berendsen paper (The Journal of Chemical Physics 81, 3684 (1984); doi: 10.1063/1.448118).</span></font></p>
<p><font size="3" face="Calibri,Arial,Helvetica,sans-serif" color="black"><span style="font-size:12pt;" id="divtagdefaultwrapper"><br>
</span></font></p>
<p><font size="3" face="Calibri,Arial,Helvetica,sans-serif" color="black"><span style="font-size:12pt;" id="divtagdefaultwrapper"><span style="background-color:white;"> "In any algorithm, you should scale the velocity at the exact same time you computed
the kinetic energy for. That is what the code since 4.6 does, right?"</span></span></font></p>
<p><font size="3" face="Calibri,Arial,Helvetica,sans-serif" color="black"><span style="font-size:12pt;" id="divtagdefaultwrapper">Yes, since 4.5. scaling is done like this.<br>
</span></font></p>
<p><font size="3" face="Calibri,Arial,Helvetica,sans-serif" color="black"><span style="font-size:12pt;" id="divtagdefaultwrapper"><br>
</span></font></p>
<p><font size="3" face="Calibri,Arial,Helvetica,sans-serif" color="black"><span style="font-size:12pt;" id="divtagdefaultwrapper">What i meant was that in this paper scaling is NOT in this way, but rather v(t+dt/2) is scaled with lambda(t-dt/2), noting:</span></font></p>
<p>"Note. Although lambda is based on the temperature at t - dt/2, its<br>
<font size="3" face="Calibri,Arial,Helvetica,sans-serif" color="black"><span style="font-size:12pt;" id="divtagdefaultwrapper"></p>
<div>value can be used to scale the velocity at t + dt/2 because of the slow variation of lambda"<font size="3" face="Calibri,Arial,Helvetica,sans-serif" color="black"><span style="font-size:12pt;" id="divtagdefaultwrapper"><br>
<br>
Thanks for your help, martin<br>
</span></font></div>
</span></font>
<p></p>
</div>
<hr style="display:inline-block;width:98%" tabindex="-1">
<div id="divRplyFwdMsg" dir="ltr"><font face="Calibri, sans-serif" style="font-size:11pt" color="#000000"><b>From:</b> gromacs.org_gmx-developers-bounces@maillist.sys.kth.se <gromacs.org_gmx-developers-bounces@maillist.sys.kth.se> on behalf of Berk Hess <hess@kth.se><br>
<b>Sent:</b> Friday, June 9, 2017 3:38:30 PM<br>
<b>To:</b> gmx-developers@gromacs.org<br>
<b>Subject:</b> Re: [gmx-developers] do_update_md</font>
<div> </div>
</div>
<div>
<div class="moz-cite-prefix">Which paper are you referring to? And what do you mean with current v?<br>
In any algorithm, you should scale the velocity at the exact same time you computed the kinetic energy for. That is what the code since 4.6 does, right?<br>
But note that as tau is usually much larger than dt, the difference will be extremely small. But the v-rescale thermostat is implemented such that one can use tau=dt, or even tau=0, and get a fully correct ensemble.<br>
<br>
Cheers,<br>
<br>
Berk<br>
<br>
On 2017-06-09 15:10, Zachmann, Martin wrote:<br>
</div>
<blockquote cite="mid:8864d1052cb24e90ae34338374586882@tum.de" type="cite"><style type="text/css" style="display:none;"><!-- P {margin-top:0;margin-bottom:0;} --></style>
<div id="divtagdefaultwrapper" style="font-size:12pt;color:#000000;font-family:Calibri,Arial,Helvetica,sans-serif;" dir="ltr">
<div>
<p>Thank you so much for your fast answers.</p>
So you are saying the difference from the code to the original paper (as is now also the case in the current master-branch<br>
in <span class="pl-en">updateMdLeapfrogSimple</span>()) is on purpose?<br>
It says in the paper that scaling the current v with a lambda calculated from minus half-step v is incorrect, yet<br>
acceptable since lambda changes slowly.<br>
regards Martin<br>
</div>
<br>
</div>
<hr style="display:inline-block;width:98%" tabindex="-1">
<div id="divRplyFwdMsg" dir="ltr"><font style="font-size:11pt" face="Calibri, sans-serif" color="#000000"><b>From:</b>
<a class="moz-txt-link-abbreviated" href="mailto:gromacs.org_gmx-developers-bounces@maillist.sys.kth.se">
gromacs.org_gmx-developers-bounces@maillist.sys.kth.se</a> <a class="moz-txt-link-rfc2396E" href="mailto:gromacs.org_gmx-developers-bounces@maillist.sys.kth.se">
<gromacs.org_gmx-developers-bounces@maillist.sys.kth.se></a> on behalf of Berk Hess
<a class="moz-txt-link-rfc2396E" href="mailto:hess@kth.se"><hess@kth.se></a><br>
<b>Sent:</b> Friday, June 9, 2017 2:51:46 PM<br>
<b>To:</b> <a class="moz-txt-link-abbreviated" href="mailto:gmx-developers@gromacs.org">
gmx-developers@gromacs.org</a><br>
<b>Subject:</b> Re: [gmx-developers] do_update_md</font>
<div> </div>
</div>
<div>
<div class="moz-cite-prefix">On 2017-06-09 14:29, Mark Abraham wrote:<br>
</div>
<blockquote cite="mid:CAMNuMARaO76EAqD7Jf=30stA4thSdfCAj+Tebhgofn5nriTj4A@mail.gmail.com" type="cite">
<div dir="ltr">Hi,
<div><br>
</div>
<div>That looks like a bug that was silently fixed during the adoption of the velocity-rescale thermostat of Bussi (see commit 333641e040ae824badf9d9531290a46ad581ed53). It may be that the issue wasn't noticed because people often used parameters that produce
lg == 1...</div>
</div>
</blockquote>
Is debatable if one should call that a bug. The original Berendsen thermostat that this code was written for did not produce the correct ensemble anyhow, so it was likely purposely coded this way to give better/more scaling with consistently too high temperature,
as would happen with the group cut-off scheme and no pair list buffering. With the V-rescale thermostat you should scale the velocity you computed the temperature for, which are the minus half-step velocities, as is now correctly done.<br>
In practice it will be difficult to detect a difference between these two codes.<br>
<blockquote cite="mid:CAMNuMARaO76EAqD7Jf=30stA4thSdfCAj+Tebhgofn5nriTj4A@mail.gmail.com" type="cite">
<div dir="ltr">
<div><br>
</div>
<div>I would strongly advise avoiding doing development work on unsupported versions of the software more than five years old... ;-) Get the git master branch and stay up to date with the hundreds of bugs that have been fixed since :-)</div>
</div>
</blockquote>
His mail didn't imply he is using an old version for development, it could just be for trying to understand the code or porting changes from an old version to master, which is good!<br>
<br>
Cheers,<br>
<br>
Berk<br>
<blockquote cite="mid:CAMNuMARaO76EAqD7Jf=30stA4thSdfCAj+Tebhgofn5nriTj4A@mail.gmail.com" type="cite">
<div dir="ltr">
<div><br>
</div>
<div>Mark</div>
</div>
<br>
<div class="gmail_quote">
<div dir="ltr">On Fri, Jun 9, 2017 at 2:01 PM Zachmann, Martin <<a moz-do-not-send="true" href="mailto:m.zachmann@tum.de">m.zachmann@tum.de</a>> wrote:<br>
</div>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">
<div dir="ltr">
<div id="m_829167863962055915divtagdefaultwrapper" style="font-size:12pt;color:#000000;font-family:Calibri,Arial,Helvetica,sans-serif" dir="ltr">
<p>Hello community!</p>
<p>i hope someone can help me understand some code changes in the do_update_md() function in update.cpp.</p>
<p>In the case where it advances r and v in the leap-frog scheme together with the berendsen thermostat<br>
</p>
<p>(last two else if { } and else { }) in version 4.0.7. it looks like this</p>
<p><span>vv = lg*(vn + f[n][d]*w_dt);<br>
</span></p>
<p><span>where lg is the lambda-scaling factor from the thermostat and w_dt is dt/m.</span></p>
<p><span>Instead in version 4.5. and higher:</span></p>
<p><span></span><span> vv = lg*vn + f[n][d]*w_dt;</span></p>
<p><span><br>
</span></p>
<p><span>The original paper (<span>The Journal of Chemical Physics 81, 3684 (1984); doi: 10.1063/1.448118</span>) seems to scale</span></p>
<p><span>both old velocity and force. Btw in <span>do_update_visc</span>() scaling is done to both.</span></p>
<p><span>What is the purpose of this?</span></p>
<p><span>any help is much appreciated,</span></p>
<p><span>regards Martin<br>
</span></p>
<p><br>
</p>
</div>
</div>
--<br>
Gromacs Developers mailing list<br>
<br>
* Please search the archive at <a moz-do-not-send="true" href="http://www.gromacs.org/Support/Mailing_Lists/GMX-developers_List" rel="noreferrer" target="_blank">
http://www.gromacs.org/Support/Mailing_Lists/GMX-developers_List</a> before posting!<br>
<br>
* Can't post? Read <a moz-do-not-send="true" href="http://www.gromacs.org/Support/Mailing_Lists" rel="noreferrer" target="_blank">
http://www.gromacs.org/Support/Mailing_Lists</a><br>
<br>
* For (un)subscribe requests visit<br>
<a moz-do-not-send="true" href="https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-developers" rel="noreferrer" target="_blank">https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-developers</a> or send a mail to
<a moz-do-not-send="true" href="mailto:gmx-developers-request@gromacs.org" target="_blank">
gmx-developers-request@gromacs.org</a>.</blockquote>
</div>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset> <br>
</blockquote>
<br>
</div>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset> <br>
</blockquote>
<br>
</div>
</body>
</html>