[gmx-users] DPPC bilayer simulation crash

[Mark Abraham]

Michael Skaug wrote:
> I am trying to perform a simulation of 128 dppc in 3655 spc water.
> I obtained .pdb and .top files from the Biocomputing website at U.Calgary.
>  I do nothing to alter the structure, which must be minimized as it comes

err, not minimized - sampled from the ensemble they were using, which 
needn't correspond to the ensemble you're using. As soon as you generate 
new velocities, the source of your structure will only give you the same 
gross morphology, not an equilibrated system.

> from the end of a 1 ns simulation.  Processing the input files works fine,
> but the mdrun always ends somewhere between step 500 and step 3000.  I have
> tried adjusting various parameters, such at the time step, tau_p, tau_t and
> the compressibility, but nothing has created a consistent run.
> I wonder where I should be focusing my time to solve this problem.  Should
> I focus on the parameters and structure, or computer hardware/software?

We can't tell unless you tell us what GROMACS is reporting as the problem.