Splitting out ifstrstream into a header and an implementation. This means that much of the code that depends on it can be compiled only once and the parallel I/O is concentrated into a few files. To do this, a number of files that relied on basic_ifstrstream to load the mpi header had to be modified to manage their own headers.
bug fixes
More changes for parallel
changes for new parallel architecture
Creating busticated version of OpenMD
Fixed a potentially serious bug in GhostBend and GhostTorsion. We were using A instead of A.transpose() where we need lab-frame coordinates of the body-fixed axis.
Fixed a bug in Torsion (and GhostTorsion) which is triggered by a configuration with colinear atoms. The problem was discovered by Brett Donovan. Thanks, Brett!
Fixed a potential bug. If Ghost atoms have no electroFrame (i.e. if they are DirectionalAtoms, but not dipoles), the GhostBend and GhostTorsions would give nans. We should be using getA instead of getElectroFrame.
Adding property set to svn entries
Fixing gradients for minimization
Fixed bugs in Restraint, refactored RestReader and RestWriter
Almost all of the changes necessary to create OpenMD out of our old project (OOPSE-4)
Added new restraint infrastructure Added MolecularRestraints Added ObjectRestraints Added RestraintStamp Updated thermodynamic integration to use ObjectRestraints Added Quaternion mathematics for twist swing decompositions Significantly updated RestWriter and RestReader to use dump-like files Added selections for x, y, and z coordinates of atoms Removed monolithic Restraints class Fixed a few bugs in gradients of Euler angles in DirectionalAtom and RigidBody Added some rotational capabilities to prinicpalAxisCalculator
many changes to keep track of particle potentials for both bonded and non-bonded interactions
Returned old LD forces to SMIPDForceManager.
A few simple mods to remove gcc compile warnings
Fixes for Inversions for use with Amber
Inversion fixes and amber mostly working
Many fixes
Changes for implementing Amber force field: Added Inversions and worked on BaseAtomTypes so that they'd function with the fortran side.
Changes required for Inversions and Base Atom types. This will break OOPSE badly for a few days or so...
Checking in changes for Hefland moment calculations
A few formatting changes to prettify the code
Removed older version of openbabel from our code. We now have a configure check to see if openbabel is installed and then we link to the stuff we need. Conversion to OOPSE's md format is handled by only one application (atom2md), so most of the work went on there. ElementsTable still needs some work to function in parallel.
Massive update to do virials (both atomic and cutoff-group) correctly. The rigid body constraint contributions had been missing and this was masked by the use of cutoff groups...
Changes to add thetacorr to dynamic props and fixes to radial rcorr.
Massive changes preparing for release of OOPSE-4: The main difference is that the new MD file format (.md, .dump, .eor) now contains meta-data information along with the configuration information.
Adding single precision capabilities to c++ side
Adding spherical boundary conditions to LD integrator
using setFrc in RigidBody::calcForcesAndTorques will discard if there are force and torque applied in rigid body itself. use addFrc instead.
fixed a bug in rigid bodies that contain directional atoms
End of the Link --> List Return of the Oject-Oriented replace yacc/lex parser with antlr parser
OptionSectionParser get compiled
streamlined reaction field for dipoles (now a good bit faster) and added reaction field for charges - still need to do charge-dipole RF
Fix the other branch of force calculation in Torision.cpp
Don't remember what we did
There is still something wrong with inertia tensor projection
erase output of inertia tensor. There is still something wrong with current implementation.
fix a bug in creating cutoffGroup. When cutoffGroup is turned off, there is a mismatch between group and center of mass array
fix a bug in projecting the inertia tensor of directional atom in rigibody into rigidbody's body frame (previous commit is wrong)
fix a bug in projecting the inertia tensor of directional atom in rigibody into rigidbody's body frame
Rotation matrix multiplication order error
putting spaceballs back where they belong
xemacs has been drafted to perform our indentation services
fix some warning messages
more memory leak are fixed
adding exclude pairs for rigidbody and cutoff group
Fixing a bug in BitSet.cpp
reactionfield get fixed
Fix for compilation on the Sun
begin bug fix
adding dynamicProps
add getRigidBodyAt function into Molecule class
adding missing implementation
adding wildcard class; NameFinder is finished
(1)Cast wrong atom to DirectionalAtom in GhostTorsion (2)get the wrong column of electroframe
(1) complete section parser's error message (2) add GhostTorsion (3) accumulate inertial tensor from the directional atoms before calculate rigidbody's inertial tensor
moved Package variable into subdirectory makefiles to speed up the building process
fix a bug in global index
port to SGI platform
merging new_design branch into OOPSE-2.0
isLinear and linearAxis are virtual in StuntDouble, but are implemented by DirectionalAtom and RigidBody In StuntDouble, they should return false and "-1" but there should be logic to figure them out in the other two classes
Shapes looks like it's working
replace old GebericData with new GenericData
change #include "mpi.h" to #include <mpi.h> in AbstractClasses.hpp
define DEV_ROOT as relative path
change Makefile to relative path
change the #include in source files
Import of OOPSE v. 2.0