LIGGGHTS precision

Submitted by marketos on
ckloss

Hi George,

generally, LIGGGHTS/LAMMPS operations are double precision. As for your question, the maximum number of digits to be included meaningfully would be something like 15 or 16, because machine epsilon is about 1e-16 for double precision

Christoph

Thu, 2 Sep 2010 - 13:14 Permalink
mkt1363

Hi Christoph,

Is it possible to increase the precision somehow? does changing to cgs helps?

Thnaks
Mohammad

Wed, 15 Mar 2017 - 18:27 Permalink

Hi,
I am creating balls and my coordinate lists have sqrt(2), sqrt(3) etc in them. I was wondering: what is the precision of LIGGGHTS? ie How many digits of the ball coordinates is it meaningful for me to include? Or, what is the furthest decimal digit that can be represented?
Thanks,
George