Open Source Software for Microscopic Calculations

Motivation and Scientific background

The advent of rapid means of exchange of software, has helped the spread of so called Open Source software, generally distributed under the GNU licensing scheme. While extremely effective in the computer science community (e.g. the Linux kernel development), and bioinformatics (report), this scheme is still not very common in the field of "atomic and molecular simulations". Nevertheless, some teams have recently started such atomic or molecular simulation open source projects:

ABINIT, a software project based on Density Functional Theory, http://www.abinit.org
MMTK, Molecular Modelling ToolKit, http://dirac.cnrs-orleans.fr/MMTK/
CAMPOS, from Density Functional to Quasi-Continuum simulation through Molecular Dynamics, http://www.fysik.dtu.dk/CAMP/CAMPOS_welcome.html
Different softwares can be accessed from http://www.ibs.fr/ext/labos/LDM/welcome_en.htm
FFTW, the Fastest Fourier Transform in the West, http://www.fftw.org
CP2K, Car-Parrinello 2000, unified Molecular Dynamics Density Functional Theory, http://cp2k.berlios.de
PWSCF and PHONON, Density-Functional Theory and Linear Response, http://www.pwscf.org
Gerhard Theurich's C++ DFT code, http://www.mrl.ucsb.edu/~theurich/Spinor
Alessandro Laio's QM/MM code, http://rgac2.ethz.ch/laio
ROBODOC, http://www.xs4all.nl/~rfsber/Robo/robodoc.html
GROMACS, molecular dynamics, http://www.gromacs.org
PyMOL, a molecular visualization package aimed at biomolecular systems, http://pymol.sourceforge.net

http://www.tc.cornell.edu/Research/Multiscale

PWPAW, http://www.wfu.edu/~natalie/papers/pwpaw/man.html
Jose-Luis Martin's pseudopotential code, http://bohr.inesc.pt/~jlm/pseudo.html
FPMD, http://www.cineca.it/~acv0/CP/carpar.html

The open source concept allows synergetic effects:

easier international collaborative work
easier cross-comparison between softwares
feed-back from users (having access to sources), e.g. bug fixes

On June 21-22, 2001, two of us organized a CECAM discussion meeting on the theme of open source softwares. During this small meeting, we had extremely lively discussions, summarized in the corresponding CECAM report. Beyond the nominal theme of the meeting, many discussions evolved around recent developments in computer sciences, that have not (or little) been used by the atomic and molecular simulation field.

In particular, it was felt that some software projects in atomic and molecular simulations became so big that it is essential to use modern software engineering concepts, such as:

libraries for handling files (HDF and NETCDF) in view of sharing and comparison of routines and codes
reusability of sources
self-documentation of codes
combination of extension and compute-intensive languages (e.g. Python+C)
high-level graphics libraries
installation and maintenance tools (autoconf, automake, CVS)
self-testing of code

These concepts are often discussed in a magazine like "Computers in Science and Engineering", that has a large diffusion. We consider it important that the developpers in our community learn about these techniques, instead of sticking to a software development style that dates from the seventies.

The motivation for this workshop is twofold:

to gather people interested in sharing their code, following the GNU spirit, and to try to establish standards that will allow them to exchange data files as well as routines
to establish contacts with the community of computer scientists who develop software tools that might be profitable to the atomic and molecular simulation community.