[gmx-users] rvdw>rlist for nve simulation

[Erik Lindahl]

Hi Linda,

This is intended to catch mistakes :-)

In this case, the mistake is that you need "switch", "shift" or 
tabulated interactions (PME, user, etc.) for energy conservation. Sharp 
cutoffs won't conserve energy, so in that case it's a waste to have 
entries in the neighborlists for which we don't calculate any 
interactions.

Gromacs also uses a very flexible shifting function, so as David 
pointed out you need to set a rvdw_shift that is smaller than rvdw.

Cheers,

Erik

On May 4, 2004, at 6:56 AM, Linda wrote:

> Hi, all,
>
> 	I found gmx3.2.1 has fixed "rlist > rcoulomb" problem when executing 
> grompp, while it still gives error message "With vdwtype = 
> Cut-off,rvdw must be >= rlist" when setting "rlist > rvdw". I want to 
> keep energy conserve when running NVE ensemble MD, so although 
> someones have pointted out one must fix it by changging code, while I 
> do not find how to fix it in mailing list archive. Can somebody help??
>
> it is OK by just simply commenting out following code in 
> src/kernel/readir.c?
>
>   if ((ir->vdwtype == evdwSWITCH) || (ir->vdwtype == evdwSHIFT)) {
>     sprintf(err_buf,"With vdwtype = %s rvdw_switch must be < rvdw",
>             evdw_names[ir->vdwtype]);
>     CHECK(ir->rvdw_switch >= ir->rvdw);
>     if (ir->rvdw_switch > ir->rvdw-0.0999) {
>       sprintf(warn_buf,"rvdw should be 0.1 to 0.3 nm larger than 
> rvdw_switch to account for diffusion and the size of charge groups");
>       warning(NULL);
>     }
>   } else {
>     sprintf(err_buf,"With vdwtype = %s,rvdw must be >= 
> rlist",evdw_names[ir->vdwtype]);
>     CHECK(ir->rlist > ir->rvdw);
>   }
>
>  	Please correct me if I am wrong. Thx!
>
> Linda
>
>
> _______________________________________________
> gmx-users mailing list
> gmx-users at gromacs.org
> http://www.gromacs.org/mailman/listinfo/gmx-users
> Please don't post (un)subscribe requests to the list. Use the
> www interface or send it to gmx-users-request at gromacs.org.