BD integrator produces -nan

GROMACS version:5.0.6
GROMACS modification: Yes
Upon scaling units properly (kBT = 1 by putting T=120…, sigma=1, D=1, gamma=1)
I get -nan from the second step. Please a bit of help here…

Hi,
I am also facing the same problem. I don’t know what is wrong. I am working on it, if get any help please post here.

I am also facing the same problem.

You don’t mention the most important paramaters: dt and bd-fric. The usual explanation is that dt/bd-fric, which is the factor converting force to displacement, is too large by one or more orders of magnitude.

Ah okay. Thank you very much. I was setting dt=1e-3 and bd-fric=1. With dt=1e-4 and bd-fric=1, it is working.

I was using dt = 0.001 and bd_fric = 0.03, but it was not working!

Yes, bd_fric should likely be on the order of 1000. Now you multiply forces of order 1000 by 0.033, which makes atoms move by tens of nm in a single step.