Thermo now can use SPRNG or rand48 (if not MPI). Finished up work on randomSPRNG. Still need to fix access to MPISimulation.
Added more code to randomSPRNG classes.
Added classes to handle random number generation.
Changed thermo getKinetic and getPotential for MPI. Added header for SPRNG random number generator class.
Added SPRNG header file.
*** empty log message ***
+ added lennard-jones force module and corresponding class. + created forceFactory directory.
*** empty log message ***
finishing work on making DumpWriter parrallel
Added mpi to InitializeFromFile read in.
Added replan of gather-scatter routines for future load balancing.
Finished updating mpiSimulation to work with c++.
*** empty log message ***
added the pareDumpLine routine to the Initialize from File class
Changed name of mpi_force_module to mpiSimulation to match with C++ code.
Initial mpi io conversion.
finished adding in the code to divide and set up the atoms and such across all of the processors.
fixed typo with #ifdef.
Added mpiComponentPlan.h a dual fortran c header file defining mpi simulation parameters for both c and fortran.
*** empty log message ***
made some shanges to streamline the setup in SimSetup
*** empty log message ***
cleaned up the divideLabour interface to use the information about Moleculestamps in the entryPlug.
*** empty log message ***
added the extraction routines so that the moleculeStamps corresponding to the actual components used in the simulation are now persistent throughout, and accessable through the SimInfo object.
Working on the clean removal of key Molecule stamps from the Hash table. stamps will be moved into a persitient linked list.
Added divideLabor.
First addition of mpiDivideLabor to split mpi simulation among processors.
*** empty log message ***
added files to divide the molecules among the processes.
Addition of c wrapper header for for fortran module proceedures.
some quick optimizations to the indexing in the storage arrays
*** empty log message ***
made some finishing changes before begining to work on the io
Initial commit of wrappers for fortran module proceedures.
*** empty log message ***
The simulation can run successfully on parrallel computers!! Now all that is need is the parrallel io routines, and parrallel forces
added simError.h to the makefile
*** empty log message ***
overhauled the makefiles to be more elegant with regard to builds. ie. now the name of the library will change when you select single or mpi build in the Makefile.inc compilers and flags will also be updated
*** empty log message ***
Fixed the whole lot to be simError safe, ands compiles. Not yet tested however
*** empty log message ***
*** empty log message ***
Bass shopuld be simError safe
the BASS event loop has finished converiosn to simError. However, it as yet untested.
continuing to refine the error checking in the BASS processing
MPI Force rewrite in progress...
Initial commit of mpi routines for force decomposition.
First commit of mpi force code.
*** empty log message ***
working on the conversion of the BASS inititialization to be parrallel safe
*** empty log message ***
adding simError into the mpiBASS Event loop
*** empty log message ***
finished updates on the easy simError change overs.
*** empty log message ***
still working on conversion over to simError
*** empty log message ***
adding in simError
*** empty log message ***
finished work on the error handling struct and functions
*** empty log message ***
started work on a global error handling scheme
*** empty log message ***
making further modifications to make FOrceField objects mpi aware
*** empty log message ***
more tinkering with the forcefield event loop
minor revisions
more minor changes to the mpi structs
finished adding in the extra structures needed to pass info through mpi. Still need to add the implementation.
finished adding the static storage arrays to the atom class
Changing internal storage of positions, torqes, forces, velocities, etc. not sure here, still working
Changing internal storage of positions, torqes, forces, velocities, etc.
*** empty log message ***
adding ForceField mpi awareness adding mpi FOrceField awareness
adding ForceField mpi awareness
*** empty log message ***
changed mpiInterface.c and .h -> mpiBASS.c and .h for clarity of purpose
*** empty log message ***
Added lhs of assignments..
Fixed double inclusion of mpiBASSEventType with define in header file.
*** empty log message ***
Clean up with icc debug flags...
*** empty log message ***
*** empty log message ***
Include makefile
quick bug fix to bassDiag
*** empty log message ***
added stamp and template diagnostic features
Added make include.
*** empty log message ***
*** empty log message ***
changed dirs...
*** empty log message ***
Initial mpi support
Addition of mpiInterface to code base.
First stab at passing base events to mpi....
begin the pain that is MPI. abandon all hope ye who check out this branch.. P.S. we've added consistent BASS error checking
fixed allot of warnings, and adde the molecule
starting work on the molecule class starting to add in the molecule class adding in the molecule class
starting work on the molecule class starting to add in the molecule class
starting work on the molecule class
fixed a bug in getA( double the_A[3][3] );
added the getA( double the_A[3][3] ) routine to the Directional Atom Class, to facilitate easier RotMat aquisition
fixed a precision error in the calculation of the temperature. Also fixed the order that dump file are written.
This commit was generated by cvs2svn to compensate for changes in r10, which included commits to RCS files with non-trunk default branches.