[XtalOpt-devel] New XtalOpt version

[David Lonie]

Hi everybody,

I've prepared a release candidate for a new version of XtalOpt. The
code is available in the "master" branch on github, and a windows
installer is here:

http://xtalopt.openmolecules.net/xtalopt/win32/RCs/XtalOpt-Avogadro-r9RC1-1.1.0-win32.exe

This update adds support for a novel method that predicts molecular
crystal packings. By performing several trial mutations and quickly
evaluating their effects on a structure's lattice energy via
forcefield calculations, the new "smart" mutation operator is expected
to produce higher quality offspring. We have successfully tested this
method on urea and benzene, finding the ground state configuration of
urea, and several polymorphs of benzene at 10 GPa (including the
ground state). The details of the algorithm and test results will be
provided in an upcoming publication.

A word of warning: This operator is new, and is in more of a
"proof-of-concept" stage. We attempted several of the molecular
structures from the most recent CCDC blind test and had difficulty
solving them without great computational expense. So for now, I'd
recommend limiting any applications to small (10-15 atom) molecules
with little/no conformational freedom that crystallize with 4 or fewer
molecules in the unit cell.

If all goes well, I'll be making an "official" release in the next few weeks.

And finally, on a personal note, I have recently taken a job with
Kitware to work on their Open Chemistry initiative
(http://www.openchemistry.org/). I'm very excited to start this work,
but this means that XtalOpt will become a "spare-time" project for me
and I'll only be making maintenance updates for the foreseeable
future. If anyone is interested in developing for XtalOpt, please
contact myself or Eva Zurek (ezurek at buffalo.edu).

Have a great summer!

Dave