Hi Sebastian, How small can I make it? I tried 0.1 but couldn't go below that value. Does it also depend on the step value for tf?
Anu On Tue, Nov 13, 2012 at 11:01 AM, Sebastian Fritsch < [email protected]> wrote: > ** > Hi Anu, > > > in case you have oscillations, it is actually more appropriate to decrease > spline_step parameter, which will make the spline 'less stiff'. > > Good luck! > > -Sebastian > > Hi Christoph and Sebastian, > > I have tried all the options with no luck. The compressibility of CG is > different than the EX so I tried adding the cg_prefactor. But this hasn't > improved the number of cycles. > > When comparing the densities like Sebastian suggested, I found that > there were oscillations so I have increased the spline length. That hasn't > helped either. The spline max right now is equal to the box length. > > I have increased ex_force_cap, played around with spline values and > equi_time. Is there anything else I can try? > > Thanks a lot! > > Anu > > On Fri, Nov 2, 2012 at 11:32 AM, Christoph Junghans <[email protected]>wrote: > >> 2012/11/2 <[email protected]>: >> > Hi Anu, >> > >> > I just have a few things to add: >> > >> > - after the first iteration, check the density profile after spline >> > smoothing in .smooth.dens (compare to the raw density file in >> .dist.new). If >> > there are oscillations, you have to tune the spline parameters. >> > - I am not sure postupd scale applies, but you can as well simply >> reduce the >> > prefactor to see if it is to high ( this may very well be if your CG >> model >> > has a different compressibility) >> Yes, postupd scale works as well, but out of all post update scripts >> scale and smooth are the only useful ones for the ft iteration. >> >> Scaling the prefactor scales the forces (before integration) while >> postupd scale scales the potential (after spline fit and integration), >> so the result could be slightly different, but in principle it does >> the same. >> >> If you add the prefactor from the coarse-grained compressibility in >> cg_prefactor, a linear interpolation between cg_prefactor and the >> prefactor from the explicit compressibility will be used as a >> prefactor in the hybrid zone automatically. >> >> Christoph >> >> > - Increase the total simulation time per iteration step and trash some >> of >> > the trajectory in the beggining using gromacs.equi_time in the settings >> file >> > >> > Good luck! >> > >> > >> > -Sebastian >> > >> > Zitat von Christoph Junghans <[email protected]>: >> > >> >> Hi, >> >> >> >> 2012/10/31 Anu Nagarajan <[email protected]>: >> >>> >> >>> Hi, >> >>> >> >>> I am trying to run the thermodynamic force using csg_inverse. I have >> the >> >>> settings.xml file with appropriate options for my system. But, when I >> try >> >>> to >> >>> run TF, I get about 2-3 iterations after which the temperature gets >> >>> really >> >>> high and the system becomes unstable. I have the latest version >> (1.2.3) >> >>> of >> >>> votca. I computed the prefactor based on the compressibility of the >> >>> atomistic region. According to the tutorial and the paper, I have set >> >>> the >> >>> TF range to be d(skin)+d(at)+d(hy)+d(skin) where d(skin) is right now >> >>> half >> >>> the cutoff distance. The entire molecule is one charge group and I >> have >> >>> 12 >> >>> atoms plus virtual site. Do I have to change any values in the >> >>> settings.xml >> >>> file? Is there anything I might be missing? >> >> >> >> The parameters will depend on the system you study. >> >> Here are some things you could try: >> >> - increase adress_ex_forcecap in the mdp file >> >> - increase the spline grid (spline_start, spline_end in >> >> cg.non-bonded.inverse.tf) in setting.xml >> >> - add a cg_prefactor in cg.non-bonded.inverse.tf, which is based on >> >> the compressibility of the cg model >> >> - scale the update (add postupd task scale) by something < 1 >> >> >> >> Sebastian, please add stuff if I have missed something. >> >> >> >> Christoph >> >>> >> >>> >> >>> I appreciate any help in this matter! >> >>> >> >>> Thanks! >> >>> Anu >> >>> >> >>> -- >> >>> You received this message because you are subscribed to the Google >> Groups >> >>> "votca" group. >> >>> To post to this group, send email to [email protected]. >> >>> To unsubscribe from this group, send email to >> >>> [email protected]. >> >>> For more options, visit this group at >> >>> http://groups.google.com/group/votca?hl=en. >> >> >> >> >> >> >> >> >> >> -- >> >> Christoph Junghans >> >> Web: http://www.compphys.de >> >> >> >> -- >> >> You received this message because you are subscribed to the Google >> Groups >> >> "votca" group. >> >> To post to this group, send email to [email protected]. >> >> To unsubscribe from this group, send email to >> >> [email protected]. >> >> For more options, visit this group at >> >> http://groups.google.com/group/votca?hl=en. >> >> >> >> >> >> -- >> >> This email was Anti Virus checked by Astaro Security Gateway. >> >> http://www.astaro.com >> > >> > >> > >> > -- >> > You received this message because you are subscribed to the Google >> Groups >> > "votca" group. >> > To post to this group, send email to [email protected]. >> > To unsubscribe from this group, send email to >> > [email protected]. >> > For more options, visit this group at >> > http://groups.google.com/group/votca?hl=en. >> > >> >> >> >> -- >> Christoph Junghans >> Web: http://www.compphys.de >> >> -- >> You received this message because you are subscribed to the Google Groups >> "votca" group. >> To post to this group, send email to [email protected]. >> To unsubscribe from this group, send email to >> [email protected]. >> For more options, visit this group at >> http://groups.google.com/group/votca?hl=en. >> >> > -- > You received this message because you are subscribed to the Google Groups > "votca" group. > To post to this group, send email to [email protected]. > To unsubscribe from this group, send email to > [email protected]. > For more options, visit this group at > http://groups.google.com/group/votca?hl=en. > > > -- > You received this message because you are subscribed to the Google Groups > "votca" group. > To post to this group, send email to [email protected]. > To unsubscribe from this group, send email to > [email protected]. > For more options, visit this group at > http://groups.google.com/group/votca?hl=en. > -- You received this message because you are subscribed to the Google Groups "votca" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/votca?hl=en.
