Files
lammps/lib
Axel Kohlmeyer 92cd2a69bf do GPU device memory accounting with doubles
using (signed) int will overflow at 2GB,
switching to unsigned has risk of hiding
overflows and using long long is not as
portable as double precisiong floating point.
2010-11-11 12:31:36 -05:00
..

This directory contains libraries that can be linked with when
building LAMMPS.  The library itself must be built first, so that a
lib*.a file exists for LAMMPS to link against.

Each library directory contains a README with additional info.  You
will need to copy one of the Makefile.* files to Makefile before
building a library.  If a Makefile.* suitable for your machine does
not exist, you will need to edit one of the existing Makefiles.

The libraries included in the LAMMPS distribution are the following:

atc           atomistic-to-continuum methods
                from Reese Jones, Jeremy Templeton, Jon Zimmerman (Sandia)
gpu	      graphical processor (GPU) routines, currently NVIDIA specific,
	        from Mike Brown (Sandia)
poems	      POEMS rigid-body integration package
                from RPI
meam	      modified embedded atom method (MEAM) potential
                from Greg Wagner (Sandia)
reax          ReaxFF potential
	        from Adri van Duin (Penn State) and Aidan Thompson (Sandia)