Re: [gmx-users] vsites and lincs-order

Berk Hess wrote: > Date: Wed, 7 Oct 2009 15:51:00 +0200 > From: joc...@xray.bmc.uu.se > To: gmx-users@gromacs.org > Subject: Re: [gmx-users] vsites and lincs-order > > All right, thanks! > > Since you mention energy conservation, maybe it would be worth adding &g

RE: [gmx-users] vsites and lincs-order

> Date: Wed, 7 Oct 2009 15:51:00 +0200 > From: joc...@xray.bmc.uu.se > To: gmx-users@gromacs.org > Subject: Re: [gmx-users] vsites and lincs-order > > All right, thanks! > > Since you mention energy conservation, maybe it would be worth adding > notes or a warnin

Re: [gmx-users] vsites and lincs-order

All right, thanks! Since you mention energy conservation, maybe it would be worth adding notes or a warnings into pdb2gmx if the time step is large. First, if lincs-order is 4 (instead of 6) and second (more important), if nstlist is not reduced with increasing dt. I strongly feel that most pe

RE: [gmx-users] vsites and lincs-order

My P-Lincs paper http://dx.doi.org/10.1021/ct700200b shows that with order 6 and a time step of 4 fs you get roughly the same constraint accuracy and energy conservation as without v-sites and a 2 fs time step. With order 4 and a 4 fs time step the energy drift is 2.2 times higher than with orde