[gmx-users] Setup Box Size
feenstra at chem.vu.nl
Thu Feb 12 10:04:08 CET 2004
Nuno R. L. Ferreira wrote:
> From: "Anton Feenstra" <feenstra at chem.vu.nl>
>>I usually take the distance between protein and box-edge slightly larger
>>the largest of rvdw or rcoul (which are always equal for me anyway), so
> for the
>>usual rvdw=rcoul=1.4 for the Gromos 43a1 ff, I use 'editconf -d 0.8 -bt
> So, the answer to my previous question posed to J. Kerrigan is: protein
> /protein distance, hence -d > 0.5*("largest of rvdw or rcoul ").
No, the distances is from protein to box-edge, which makes the protein-
protein distance twice that. A simple '1d pbc' example may clarify:
| PPP | QQQ |
'PPP' is the 'real' protein, 'QQQ' one of its images. '|' is the box edge.
Setting '-d' to 0.8, will keep the box 0.8 nm away from the protein, and
so the 'PPP'-'QQQ' distance will be about 1.6nm, larger than the cutoff.
> BTW, why so many people talk about the advantages of using PME, but still
> use coulombtype = cuttof?
> Time computational cost? Bad scalling in parallel runs?
> I assume that rvdw=rcoul=1.4 means that you're doing cutoff for vdw and
Yes it does. The reason is simple: I've always used it up to now, and
it works for me. I'm considering switching to PME, but for the running
work I like to stay consistent with previous simulations. Actually, I
just started some pilot runs on a (heavily charged) peptide with PME.
Now I'll have to learn how to add counter-ions... ;-)
| | |
| _ _ ___,| K. Anton Feenstra |
| / \ / \'| | | Dept. of Pharmacochem. - Vrije Universiteit Amsterdam |
|( | )| | | De Boelelaan 1083 - 1081 HV Amsterdam - Netherlands |
| \_/ \_/ | | | Tel: +31 20 44 47608 - Fax: +31 20 44 47610 |
| | Feenstra at chem.vu.nl - www.chem.vu.nl/~feenstra/ |
| | "If You See Me Getting High, Knock Me Down" |
| | (Red Hot Chili Peppers) |
More information about the gromacs.org_gmx-users