<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html><head><title></title>
<META http-equiv=Content-Type content="text/html; charset=utf-8">
<meta http-equiv="Content-Style-Type" content="text/css">
<style type="text/css"><!--
body {
margin: 5px 5px 5px 5px;
background-color: #ffffff;
}
/* ========== Text Styles ========== */
hr { color: #000000}
body, table /* Normal text */
{
font-size: 9pt;
font-family: 'Courier New';
font-style: normal;
font-weight: normal;
color: #000000;
text-decoration: none;
}
span.rvts1 /* Heading */
{
font-size: 10pt;
font-family: 'Arial';
font-weight: bold;
color: #0000ff;
}
span.rvts2 /* Subheading */
{
font-size: 10pt;
font-family: 'Arial';
font-weight: bold;
color: #000080;
}
span.rvts3 /* Keywords */
{
font-size: 10pt;
font-family: 'Arial';
font-style: italic;
color: #800000;
}
a.rvts4, span.rvts4 /* Jump 1 */
{
font-size: 10pt;
font-family: 'Arial';
color: #008000;
text-decoration: underline;
}
a.rvts5, span.rvts5 /* Jump 2 */
{
font-size: 10pt;
font-family: 'Arial';
color: #008000;
text-decoration: underline;
}
span.rvts6
{
font-size: 11pt;
font-family: 'segoe ui';
font-weight: bold;
color: #ffffff;
background-color: #0000ff;
}
span.rvts7
{
font-size: 11pt;
font-family: 'segoe ui';
}
a.rvts8, span.rvts8
{
font-size: 11pt;
font-family: 'segoe ui';
color: #0000ff;
text-decoration: underline;
}
span.rvts9
{
font-size: 8pt;
font-family: 'arial';
font-style: italic;
color: #c0c0c0;
}
a.rvts10, span.rvts10
{
font-size: 8pt;
font-family: 'arial';
color: #0000ff;
text-decoration: underline;
}
/* ========== Para Styles ========== */
p,ul,ol /* Paragraph Style */
{
text-align: left;
text-indent: 0px;
padding: 0px 0px 0px 0px;
margin: 0px 0px 0px 0px;
}
.rvps1 /* Centered */
{
text-align: center;
}
--></style>
</head>
<body>
<p>Hang,</p>
<p><br></p>
<p>This is exactly the problem. Your system isn't set up properly, and you are trying to mask this fact behind constraints. By the way, oscillatory behavior does not necessarily mean the system is too soft. It may actually mean that your set of connstants is causing resonant or near-resonant oscillations, given the time steps you've got. For instance, it is good practice to check if the coarse-grain time step corresponds to the ratio between the bead mass and the average per-particle mass of all atoms assigned to a bead. In the case of homogeneous crystals, it is merely the number of atoms in one bead. If the number of atoms in one bead is five and you're hiking up the time step from 1 ps to 20 ps, there's likely going to be a problem. </p>
<p><br></p>
<p>It is good that LINCS is screaming, it would be far more dangerous if it didn't.</p>
<p><br></p>
<p>Diamond (as a crystal type) is one of the most stable ones, and wild oscillations are usually indicative of the setup being wrong. There are of course ways to try making LINCS/SHAKE work in this situation, but that's just bad science, in my opinion.</p>
<p><br></p>
<p>Good luck!</p>
<p>
<br>Alex</p>
<p><br></p>
<p><br></p>
<div><table border=0 cellpadding=1 cellspacing=2>
<tr valign=top>
<td width=11 style="background-color: #0000ff;">
<p><span class=rvts6>></span></p>
</td>
<td width=1215 style="background-color: #ffffff;">
<p><span class=rvts7>Hi Alex,</span></p>
<p><span class=rvts7>When I ran my diamond with bonds and angles interactions, the beads in the diamond oscillated significantly, larger time step larger amplitude. The crystal shape could not conserved.</span></p>
<p><span class=rvts7>Tomorrow I'll try SHAKE.</span></p>
<p><span class=rvts7>Really thanks for your help! It is deep night here, so I may reply you on later morning. :)</span></p>
<p><span class=rvts7>Hang</span></p>
<p><span class=rvts7>2015年4月21日 上午12:49于 "Alex" <</span><a class=rvts8 href="mailto:nedomacho@gmail.com">nedomacho@gmail.com</a><span class=rvts7>>写道：</span></p>
<p><br></p>
<p><span class=rvts7>Hi Hang,</span></p>
<p><span class=rvts7>Are the angles in the system conserved via artificial bonds? In other words, when you run your diamond, does it lose crystal shape, or does it stay diamond? I am just trying to understand what you call "soft". You can always try SHAKE instead of LINCS, if you insist on constraints, i'm just trying to see if these constraints are masking something really wrong with the unconstrained setup. You see, noone uses constraints in solid-state simulations, the structure must be conserved, unless something is seriously wrong.</span></p>
<p><span class=rvts7>Alex</span></p>
<p><span class=rvts7>On Apr 20, 2015 10:34 AM, "hang yin" <</span><a class=rvts8 href="mailto:yinhang503@gmail.com">yinhang503@gmail.com</a><span class=rvts7>> wrote:</span></p>
<p><span class=rvts7>></span></p>
<p><span class=rvts7>> Hi Alex,</span></p>
<p><span class=rvts7>></span></p>
<p><span class=rvts7>> Thanks for your reply. </span></p>
<p><span class=rvts7>></span></p>
<p><span class=rvts7>> I used large time step because it is the default value(20 ~40 ps) for the Martini Coarse Grained ff. Applied that large time step, the angles and bonds have been proved that they could not keep ND rigid enough. And for a very large system I want to build in future, the CG model have to be used, meaning that I have to solve the rigidity problem in a large time step setup. Hence I really want to use constraints here. </span></p>
<p><span class=rvts7>></span></p>
<p><span class=rvts7>> Hang</span></p>
<p><span class=rvts7>></span></p>
<p><span class=rvts7>></span></p>
<p><span class=rvts7>> 2015-04-21 0:22 GMT+08:00 Alex <</span><a class=rvts8 href="mailto:nedomacho@gmail.com">nedomacho@gmail.com</a><span class=rvts7>>:</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> Also, I just looked at your topology. Where are your angles? Try to</span></p>
<p><span class=rvts7>>> implement them and then turn off constraints.</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> Alex</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> A> Is there a particular reason you're applying LINCS constraints to your</span></p>
<p><span class=rvts7>>> A> diamond structure? Usually, LINCS gives convergence</span></p>
<p><span class=rvts7>>> A> errors/warnings if the system is poorly</span></p>
<p><span class=rvts7>>> A> equilibrated and/or your time step is too large. The fact that your</span></p>
<p><span class=rvts7>>> A> system was "soft" without constraints actually suggests poor setup,</span></p>
<p><span class=rvts7>>> A> which once again could include the time step.</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> A> Diamond structure, if properly set up, should be quite rigid, even if</span></p>
<p><span class=rvts7>>> A> your bond/angle terms are somewhat off, because, well, it's the</span></p>
<p><span class=rvts7>>> A> diamond structure.</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> A> Alex</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> Hi all,</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> My research focuses on the dynamics of a nanodiamond(ND) in a biological</span></p>
<p><span class=rvts7>>> hy>> environment. I am working with gromacs version 5.0.2 and using Martini CG</span></p>
<p><span class=rvts7>>> hy>> ff. In order to keep the rigidity of ND, I applied constraints as the stiff</span></p>
<p><span class=rvts7>>> hy>> bond in the .itp file. But my system crashed in the NVT run with LINCS</span></p>
<p><span class=rvts7>>> hy>> WARNING. So I tried to use a small ND, an octahedron, in the water box for</span></p>
<p><span class=rvts7>>> hy>> debugging. It did work well. However, after I increased its size(just add 3</span></p>
<p><span class=rvts7>>> hy>> beads actually), it crashed with same input parameters.</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> How it could happen? I am the beginner using gromacs and CG model. I think</span></p>
<p><span class=rvts7>>> hy>> the problem is due to the constraints because the same system with defined</span></p>
<p><span class=rvts7>>> hy>> bond interaction worked well(but the ND in this situation is soft). I am</span></p>
<p><span class=rvts7>>> hy>> just wondering that is there any limitation for applying constraints for</span></p>
<p><span class=rvts7>>> hy>> crystal packing system or anything wrong with my configuration?</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> Seeking for help. Thanks a lot.</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> Kevin</span></p>
<p><span class=rvts7>>> hy>> ******************************************************************************</span></p>
<p><span class=rvts7>>> hy>> My itp file:</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> ;itp file of ND</span></p>
<p><span class=rvts7>>> hy>> [ moleculetype ]</span></p>
<p><span class=rvts7>>> hy>> ; molname nrexcl</span></p>
<p><span class=rvts7>>> hy>> CG-ND 1</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> [ atoms ]</span></p>
<p><span class=rvts7>>> hy>> ; id type resnr residu atom cgnr charge</span></p>
<p><span class=rvts7>>> hy>> 1 P1 1 FO CD1 1 0.000000</span></p>
<p><span class=rvts7>>> hy>> 2 C1 1 FH CD2 2 0.000000</span></p>
<p><span class=rvts7>>> hy>> 3 P1 1 FO CD3 3 0.000000</span></p>
<p><span class=rvts7>>> hy>> 4 C1 1 FH CD4 4 0.000000</span></p>
<p><span class=rvts7>>> hy>> 5 C1 2 FH CD1 5 0.000000</span></p>
<p><span class=rvts7>>> hy>> 6 C1 2 FH CD2 6 0.000000</span></p>
<p><span class=rvts7>>> hy>> 7 C1 2 FH CD3 7 0.000000</span></p>
<p><span class=rvts7>>> hy>> 8 P1 2 FO CD4 8 0.000000</span></p>
<p><span class=rvts7>>> hy>> 9 P1 3 FO CD1 9 0.000000</span></p>
<p><span class=rvts7>>> hy>> 10 P1 3 FO CD2 10 0.000000</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> [ constraints ]</span></p>
<p><span class=rvts7>>> hy>> ; i j funct length force.c.</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> #ifdef FLEXIBLE</span></p>
<p><span class=rvts7>>> hy>> [ bonds ]</span></p>
<p><span class=rvts7>>> hy>> #endif</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> 1 2 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 1 3 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 1 4 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 2 3 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 2 4 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 3 4 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 3 5 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 3 6 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 4 5 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 4 6 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 5 6 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 5 7 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 5 8 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 6 7 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 6 8 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 7 10 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 7 8 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 7 9 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 8 10 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 8 9 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> 9 10 1 0.317 1000000</span></p>
<p><span class=rvts7>>> hy>> ******************************************************************************</span></p>
<p><span class=rvts7>>> hy>> My mdp file:</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> title = debugging for ND</span></p>
<p><span class=rvts7>>> hy>> ; Run parameters</span></p>
<p><span class=rvts7>>> hy>> integrator = md ; leap-frog integrator</span></p>
<p><span class=rvts7>>> hy>> nsteps = 2500 ; 20 * 2500 = 500 ps (0.05 ns)</span></p>
<p><span class=rvts7>>> hy>> dt = 0.02 ; 20 fs</span></p>
<p><span class=rvts7>>> hy>> ; Output control</span></p>
<p><span class=rvts7>>> hy>> nstxout = 1 ; save coordinates</span></p>
<p><span class=rvts7>>> hy>> nstvout = 1 ; save velocities</span></p>
<p><span class=rvts7>>> hy>> nstenergy = 10 ; save energies</span></p>
<p><span class=rvts7>>> hy>> nstlog = 10 ; update log file</span></p>
<p><span class=rvts7>>> hy>> nstcomm = 10</span></p>
<p><span class=rvts7>>> hy>> fcstep = 0</span></p>
<p><span class=rvts7>>> hy>> nstcalcenergy = 10</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> ; Bond parameters</span></p>
<p><span class=rvts7>>> hy>> unconstrained_start = no</span></p>
<p><span class=rvts7>>> hy>> constraints = none</span></p>
<p><span class=rvts7>>> hy>> constraint_algorithm = Lincs</span></p>
<p><span class=rvts7>>> hy>> lincs_order = 4</span></p>
<p><span class=rvts7>>> hy>> lincs_warnangle = 30</span></p>
<p><span class=rvts7>>> hy>> lincs-iter = 1</span></p>
<p><span class=rvts7>>> hy>> morse = no</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> ; Neighborsearching</span></p>
<p><span class=rvts7>>> hy>> cutoff-scheme = verlet</span></p>
<p><span class=rvts7>>> hy>> verlet-buffer-drift = 0.001</span></p>
<p><span class=rvts7>>> hy>> ns_type = grid ; search neighboring grid cels</span></p>
<p><span class=rvts7>>> hy>> nstlist = 10 ;</span></p>
<p><span class=rvts7>>> hy>> rlist = 1.4 ; short-range neighborlist cutoff (in nm)</span></p>
<p><span class=rvts7>>> hy>> rcoulomb = 1.2 ; short-range electrostatic cutoff (in nm)</span></p>
<p><span class=rvts7>>> hy>> rvdw = 1.2 ; short-range van der Waals cutoff (in nm)</span></p>
<p><span class=rvts7>>> hy>> vdw_type = cutoff</span></p>
<p><span class=rvts7>>> hy>> rvdw_switch = 0.9</span></p>
<p><span class=rvts7>>> hy>> vdw-modifier = Potential-shift</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> ; Electrostatics</span></p>
<p><span class=rvts7>>> hy>> coulombtype = Reaction_field</span></p>
<p><span class=rvts7>>> hy>> coulomb_modifier= Potential-shift</span></p>
<p><span class=rvts7>>> hy>> rcoulomb_switch = 0.0</span></p>
<p><span class=rvts7>>> hy>> epsilon_r = 15</span></p>
<p><span class=rvts7>>> hy>> epsilon_rf = 0</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> ; Temperature coupling is on</span></p>
<p><span class=rvts7>>> hy>> tcoupl = Berendsen</span></p>
<p><span class=rvts7>>> hy>> tc-grps = system ; three coupling groups - more accurate</span></p>
<p><span class=rvts7>>> hy>> tau_t = 1.0 ; time constant, in ps</span></p>
<p><span class=rvts7>>> hy>> ref_t = 310 ; reference temperature, one for each group, in K</span></p>
<p><span class=rvts7>>> hy>> ; Pressure coupling is on</span></p>
<p><span class=rvts7>>> hy>> pcoupl = no ; Pressure coupling on in NPT</span></p>
<p><span class=rvts7>>> hy>> pcoupltype = isotropic ; uniform scaling of x-y-z box vectors</span></p>
<p><span class=rvts7>>> hy>> tau_p = 4.0 ; time constant</span></p>
<p><span class=rvts7>>> hy>> ref_p = 1.0 ; reference pressure</span></p>
<p><span class=rvts7>>> hy>> compressibility = 1e-5 ; isothermal compressibility, bar^-1</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> ; Periodic boundary conditions</span></p>
<p><span class=rvts7>>> hy>> pbc = xyz ; 3-D PBC</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> hy>> ; Velocity generation</span></p>
<p><span class=rvts7>>> hy>> gen_vel = no</span></p>
<p><span class=rvts7>>> hy>> gen_temp = 310</span></p>
<p><span class=rvts7>>> hy>> gen_seed = -1</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> --</span></p>
<p><span class=rvts7>>> Best regards,</span></p>
<p><span class=rvts7>>> Alex mailto:</span><a class=rvts8 href="mailto:nedomacho@gmail.com">nedomacho@gmail.com</a></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> --</span></p>
<p><span class=rvts7>>> Gromacs Users mailing list</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> * Please search the archive at</span><a class=rvts8 href="http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List"> http://www.gromacs.org/Support/Mailing_Lists/GMX-Users_List</a><span class=rvts7> before posting!</span></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> * Can't post? Read</span><a class=rvts8 href="http://www.gromacs.org/Support/Mailing_Lists"> http://www.gromacs.org/Support/Mailing_Lists</a></p>
<p><span class=rvts7>>></span></p>
<p><span class=rvts7>>> * For (un)subscribe requests visit</span></p>
<p><span class=rvts7>>></span><a class=rvts8 href="https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users"> https://maillist.sys.kth.se/mailman/listinfo/gromacs.org_gmx-users</a><span class=rvts7> or send a mail to</span><a class=rvts8 href="mailto:gmx-users-request@gromacs.org"> gmx-users-request@gromacs.org</a><span class=rvts7>.</span></p>
<p><span class=rvts7>></span></p>
<p><span class=rvts7>></span></p>
</td>
</tr>
</table>
</div>
<p><br></p>
<p><br></p>
<p><br></p>
<p><br></p>
<p><span class=rvts9>-- </span></p>
<p><span class=rvts9>Best regards,</span></p>
<p><span class=rvts9> Alex </span><a class=rvts10 href="mailto:nedomacho@gmail.com">mailto:nedomacho@gmail.com</a></p>
</body></html>