[gmx-users] Gromacs 3.1 beta released

[Erik Lindahl]

Hi everybody,

We've just released the beta version of Gromacs 3.1 in case you're
interested in trying it out. It's mostly a maintenance relase, but it
does contain some new and useful features, including:

* Tons of bugfixes and minor extensions
* Arbitrary subsets of the system can be temperature-coupled while other
parts of the system aren't.
* Successive overrelaxation for Shake.
* nmrun & xmdrun have been merged with mdrun. Normal modes is just a
specific integrator now.
* Conjugate gradients work better with PME.
* A bug in the charge and dipole correction code has been fixed.
* New analysis tools to calculate cluster sizes
* Altivec assembly loops on PowerPC G4 processors, both for Linux and
Mac OS X.
* SSE2 double precision assembly loops on Pentium 4 processors.
* Innerloops than only calculate potential, not force. These save 30-50%
runtime if you are using Gromacs as a library to perform Monte-Carlo
simulations. We'll try to add a general extensible Monte Carlo
integrator in the future, too.
* The intel compilers are supported both for ia-32 (x86) and ia-64
(itanium) systems.
* You can access the gromacs CVS repository read only now! Check out the
developer section of www.gromacs.org.
* Many more systems & compilers supported by default


You can find the source code and rpm packages at
ftp://ftp.gromacs.org/pub/beta/

Althought it's formally a beta, we expect it to be more stable than the
current released versions, and it can read all your old files.

Trajectories haven't changed at all, while the energy files are
identical unless you use position restraints, and the .tpr files have a
couple of new options in them. Gromacs 3.1 can of course read all your
3.0 files, though, and unless something extreme happens the format won't
change for the final release.

Please report bugs to gmx-users at gromacs.org, or gromacs at gromacs.org. If
no serious bugs appear we plan to make the real release around march
1st.

I think we've fixed all the problems we've seen on the list, but it's
possible that we've missed something since we don't have an automatic
bugtracking system (yet)....


And, by the way, a GREAT thanks to all of you - we've found & corrected
ten times more bugs with your help on this list than we would have been
able to do on our own! I'm sure there are still things we have
overlooked, though, so go ahead, try it out, and post your comments
before the real release ;-)

Cheers,

Erik