Hello Justin,<br>Thanks for your reply.<br><br><pre>Yes, absolutely. The Berendsen method leads to faster convergence, while <br>Parrinello-Rahman often fluctuates to a larger extent before settling around the <br>target pressure values. It is not suitable for equilibration, generally. Have <br>
you done prior equilibration for this system?</pre>
Actually, this run is an equilibration. So I guess for pressure coupling, it is good to use Berendsen for equilibration rather than PR?<br><br><pre>Also note that you're not providing several important parameters in your <br>
pressure coupling section, like reference pressures and compressibilities. With <br>anisotropic pressure coupling, you should make sure that the off-diagonal <br>components are not causing compression or distortion of your system, as the <br>
manual advises.</pre>
Yes, these parameters are there in the .mdp file (as shown below), in the previous email, I didn't include them. <br>pcoupl = Parrinello-Rahman<br>pcoupltype = anisotropic<br>nstpcouple = -1<br>
tau_p = 1.0<br>compressibility = 4.6e-5 4.6e-5 4.6e-5 0 0 0<br>ref_p = 1.0 1.0 1.0 0.0 0.0 0.0<br><br>thanks<br>Shyno<br><br clear="all"><pre>Shyno Mathew wrote:<br>><i> Dear gromacs users,<br>
</i>><i> I am having some difficulty in setting up an NPT simulation, following <br></i>><i> are the issues<br></i>><i> <br></i>><i> 1. With Parinello-Rahman coupling, I am getting the error:<br></i>><i> Fatal error:<br>
</i>><i> The X-size of the box (4.224258) times the triclinic skew factor <br></i>><i> (1.000000) is smaller than the number of DD cells (5) times the smallest <br></i>><i> allowed cell size (0.844801)<br></i>><i> As mentioned in the archives, initially I thought my system is too small <br>
</i>><i> to split between different processors. When I load the trajectory I can <br></i>><i> see that the box is shrinking very fast in the x direction.<br></i><br>It may be possible that your system can be split over the chosen number of <br>
processors, but the margin of error is very small. If the system is not well <br>equilibrated, the box may be changing quite fast.<br><br>><i> I have performed the same run with Berendsen thermostat and it ran <br></i>><i> fine. So changing the pcoupling method can cause such a change?<br>
</i><br>Yes, absolutely. The Berendsen method leads to faster convergence, while <br>Parrinello-Rahman often fluctuates to a larger extent before settling around the <br>target pressure values. It is not suitable for equilibration, generally. Have <br>
you done prior equilibration for this system?<br><br>><i> With the berendsen p coupling I have done a benchmark for the system and <br></i>><i> I am using the optimum number of processors. Will the change in <br></i>><i> pcoupling affects benchmark as well?<br>
</i>><i> <br></i><br>The differences should be minimal. I doubt there is a significant difference in <br>calculation times for the various pressure coupling algorithms, at least when <br>considering the other truly time-consuming processes in the MD algorithm.<br>
<br>><i> some parameters in .mdp file:<br></i>><i> <br></i>><i> integrator = sd<br></i>><i> pbc = xyz<br></i>><i> rlist = 1.2<br></i>><i> coulombtype = PME<br>
</i>><i> fourierspacing = 0.12<br></i>><i> pme_order = 4<br></i>><i> ewald_rtol = 1e-05<br></i>><i> optimize_fft = no<br></i>><i> epsilon_surface = 0<br>
</i>><i> ewald_geometry = 3d<br></i>><i> rcoulomb = 1.2<br></i>><i> vdwtype = Shift<br></i>><i> rvdw_switch = 0.9<br></i>><i> rvdw = 1.0<br>
</i>><i> epsilon_r = 1<br></i>><i> DispCorr = EnerPres<br></i>><i> tc-grps = system<br></i>><i> tau_t = 2.0<br></i>><i> ref_t = 310<br>
</i>><i> pcoupl = Parrinello-Rahman<br></i>><i> pcoupltype = anisotropic<br></i>><i> nstpcouple = -1<br></i>><i> tau_p = 1.0<br></i>><i> <br>
</i>><i> 2. I changed some of the above mentioned parameters, for example vdwtype <br></i>><i> to cut-off and DispCorr to energy as shown below:<br></i>><i> <br></i>><i> rlist = 1.0<br></i>><i> rcoulomb = 1.0<br>
</i>><i> vdwtype = Cut-off<br></i>><i> rvdw = 1.0<br></i>><i> epsilon_r = 1<br></i>><i> DispCorr = Ener<br></i>><i> <br></i>><i> Again I get the same error:<br>
</i>><i> Fatal error:<br></i>><i> The X-size of the box (5.000070) times the triclinic skew factor <br></i>><i> (1.000000) is smaller than the number of DD cells (5) times the smallest <br></i>><i> allowed cell size (1.000000)<br>
</i>><i> <br></i><br>The fact that this error persists despite other changes suggests to me that <br>there is simply inherently wrong with the system, i.e. inadequate minimization <br>or equilibration, unstable topology, etc.<br>
<br>Also note that you're not providing several important parameters in your <br>pressure coupling section, like reference pressures and compressibilities. With <br>anisotropic pressure coupling, you should make sure that the off-diagonal <br>
components are not causing compression or distortion of your system, as the <br>manual advises.<br><br><br>><i> Also I thought for obtaining more accurate trajectory, I should apply <br></i>><i> dispersion corrections for both pressure and energy, is this true?<br>
</i>><i> <br></i><br>With a plain cutoff, dispersion correction is often advisable.<br><br>><i> 3. In the mailing list I found that tau_p has to be greater than tau_t <br></i>><i> to avoid larger fluctuations and so changed the parameters as shown <br>
</i>><i> below, everything remains same as in question2 except the following:<br></i>><i> <br></i>><i> tau_t = 2.0<br></i>><i> ref_t = 310<br></i>><i> tau_p = 3.0<br>
</i>><i> <br></i><br>When citing information found in the archive, providing a link is advisable. I <br>won't categorically say whether that advice is right or wrong, though it is <br>generally correct given the nature of barostats vs. thermostats. Do not <br>
necessarily believe that tau_p > tau_t will lead to greater stability. The <br>simple fact that pressure oscillates more widely than temperature is the main <br>reason for this need, it's simply fundamental.<br><br>
><i> <br></i>><i> Fatal error:<br></i>><i> One of the box vectors has become shorter than twice the cut-off length <br></i>><i> or box_yy-|box_zy| or box_zz has become smaller than the cut-off.<br></i>><i> <br>
</i><br>This is the same problem as before, but by chance a different error was <br>triggered based on the algorithm that failed first. Your system is blowing up. <br> See the explanation and advice here:<br><br><a href="http://www.gromacs.org/Documentation/Terminology/Blowing_Up">http://www.gromacs.org/Documentation/Terminology/Blowing_Up</a><br>
<br>-Justin<br><br>-- <br>========================================<br><br>Justin A. Lemkul<br>Ph.D. Candidate<br>ICTAS Doctoral Scholar<br>MILES-IGERT Trainee<br>Department of Biochemistry<br>Virginia Tech<br>Blacksburg, VA<br>
jalemkul[at]<a href="http://vt.edu">vt.edu</a> | (540) 231-9080<br><a href="http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin">http://www.bevanlab.biochem.vt.edu/Pages/Personal/justin</a><br></pre><br>-- <br>Shyno Mathew<br>
PhD student<br>Department of Chemical Engineering<br>Columbia University<br><br>