Re: [AMBER] Fwd: GPU pmemd.cuda runs very slow (3ns/day)???

From: Jason Swails <>
Date: Thu, 4 Jun 2015 08:51:11 -0400

On Thu, Jun 4, 2015 at 8:35 AM, Patrick McGrath <> wrote:

> Hi Jason,
> Would this cutoff also appropriate for membrane simulations?

​Depends on who you ask and what membrane force field you're using. If
you're using the CHARMM force field, they'll tell you that you need a 12 A
cutoff and ideally a vdW force-switch turned on at 10 A (which Amber
doesn't have).

The Lipid14 paper uses a 10 A cutoff, so I would suggest using that for
Lipid14 simulations. The long-range dispersion correction assumes
isotropic particle distribution, and membranes are anisotropic systems. As
a result, the correction is less accurate for lipid simulations, so a
larger cutoff is probably justified.

All the best,

Jason M. Swails
Rutgers University
Postdoctoral Researcher
AMBER mailing list
Received on Thu Jun 04 2015 - 06:00:03 PDT
Custom Search