Re: Rule of thumb for choosing time step in MD

From: David Hardy (
Date: Tue Mar 16 2021 - 16:23:21 CDT

Dear Dr. Mi,

Are you talking about energy fluctuations? The energy is expected to fluctuate.

There are known stability limits on the step size. Specifically, the time step should be no more than about 1/5 the period of the fastest oscillations present in the system (see <>). This is the big motivation for constraining bonds to hydrogen atoms (“rigidBonds all”) in order to eliminate what would otherwise be the fastest oscillations, permitting use of multiple time stepping with a base time step of up to 2 fs and a longest time step of 4 fs for the long-range electrostatics. Simulating without constraints, you probably shouldn’t exceed a base time step of 1 fs and a longest time step of 2 fs for the long-range electrostatics.

Best regards,

David J. Hardy, Ph.D.
Beckman Institute
University of Illinois at Urbana-Champaign
405 N. Mathews Ave., Urbana, IL 61801,
> On Mar 16, 2021, at 3:12 PM, Mi Yang <> wrote:
> Dear Colleagues,
> What is the best time step for getting less fluctuations during NAMD simulations?
> Is there any rule of thumb for Protein alone or Ligand-protein complex inputs?
> Any comment will be highly appreciating!
> Dr. Mi

This archive was generated by hypermail 2.1.6 : Fri Dec 31 2021 - 23:17:11 CST