Date: Mon, 26 Aug 2002 18:23:52 -0700
From: "David A. Case"
Subject: Re: ewald bomb in md equilibration (fwd)

On Mon, Aug 26, 2002, amber@heimdal.compchem.ucsf.edu wrote:
>
>
> I am trying to do an MD run on my ligand/receptor system. The system was first
> equilibrated under the NVT ensemble with no problems. However, when trying to
> do the MD run under the NPT ensemble,
> irest = 1, ntx = 5,
> ntb = 2, ntp = 1,
> nstlim = 100000,
> dt = 0.0005, vlimit = 0.0,
> ntpr = 1000, tautp = 0.2, taup = 0.4,
> iwrap=1, comp = 108.6, npscal = 1,
> ntt = 4,
> nsnb = 1, vrand = 1000, nscm = 500,
> temp0 = 300.,
> cut = 14.0,
> scee = 1.2,
> ntf = 2,
> ntc = 2,
> the following error message comes up:
>
> EWALD BOMB in subroutine ewald_list
> volume of ucell too big, too many subcells
> list grid memory needs to be reallocated, restart sander
>
> How do I fix the problem?
>

You could restart sander, as suggested. But try using a much smaller cutoff
(say the default of 8.0, or 9.0 if you are more conservative). I would set
ntt=1 and taup to a larger value (say 2.0). I would also suggest running
a shorter simulation, and setting ntpr to a small value so that you can
monitor what is going on better.

..hope this helps...dac