This isn't surprising. Pressure coupling without explicit solvent
doesn't work the way that you are anticipating. You might want to look
at what happened to the system using ngmx or VMD to see what went wrong.
Unless you need to simulate in a box for some reason, you should turn
off PBC as Justin
Scratch what I said last night about pressure coupling - it crashes
after ~200K timesteps due to fluctuations in the cell volume. I'm
sticking with NVT for now.
On 5/31/2011 6:14 PM, Michael Daily wrote:
The following mdp file produces a successful dynamics run out to 100K
steps / 200 ps. Wh
Thanks for the suggestions. I can see how a smaller dt would probably be
the most general approach to use as it should work with just about any
reasonable combination of settings.
-- Josh
On Tue, 2011-05-31 at 11:02 +1000, Mark Abraham wrote:
> On 31/05/2011 10:54 AM, Justin A. Lemkul wrote:
> >
The following mdp file produces a successful dynamics run out to 100K
steps / 200 ps. What I discovered is this:
Using the md integrator, it is necessary to turn off pressure coupling.
However, pressure coupling works with sd (Langevin) integrator.
Mike
---
; title and include files
title
Thanks for the recommendations everyone. I tried all of the mdp changes
recommended by Justin (increase rlist, rvdw, etc to 2.0; change
T-coupling to v-scale, and eliminate P-coupling). When I increased the
distance cutoffs, it ran about 30 ps then crashed instead of crashing
immediately. On
On 31/05/2011 10:54 AM, Justin A. Lemkul wrote:
Joshua L. Phillips wrote:
I've found that I often get LINCS warnings like this when starting from
highly extended conformations when using implicit solvent. The GBSA
surface tension combined with the lack of viscosity (due to the absence
of expli
On 31/05/2011 8:50 AM, Justin A. Lemkul wrote:
One option that might be advantageous is to use the all-vs-all kernels
for a speed upgrade. You can accomplish this with:
rlist = 0
nstlist = 0
rvdw = 0
rcoulomb = 0
rgbradii = 0
pbc = no
comm-mode = angular
You'd have to run with mdrun -pd (part
Joshua L. Phillips wrote:
I've found that I often get LINCS warnings like this when starting from
highly extended conformations when using implicit solvent. The GBSA
surface tension combined with the lack of viscosity (due to the absence
of explicit water) allows the protein to change conformat
I've found that I often get LINCS warnings like this when starting from
highly extended conformations when using implicit solvent. The GBSA
surface tension combined with the lack of viscosity (due to the absence
of explicit water) allows the protein to change conformation much faster
than LINCS lik
Thanks Justin, this is very helpful. I'll attempt these fixes tomorrow.
Mike
On 5/30/2011 5:50 PM, Justin A. Lemkul wrote:
Michael D. Daily wrote:
Hi all,
I'm trying to run implicit solvent calculations in gromacs 4.5 with
the charmm forcefield. I am able to minimize successfully and
c
Michael D. Daily wrote:
Hi all,
I'm trying to run implicit solvent calculations in gromacs 4.5 with the
charmm forcefield. I am able to minimize successfully and compile for
When troubleshooting, it is always advisable to try the latest version (4.5.4)
to see if the problem is reproduci
Hi all,
I'm trying to run implicit solvent calculations in gromacs 4.5 with the
charmm forcefield. I am able to minimize successfully and compile for
mdrun, but soon after starting, mdrun complains about excessive rotation
in LINCS (see the error printed below that). I also include my mdp fi
12 matches
Mail list logo
