CVS log for tahoe/src/main/FEExecutionManagerT.cpp

Added references, cleared up some comments, cleaned up leftovers from move of BSM code

merging changes from BRANCH_david_spatial_2

external I/O constructed inside FEManagerT

emit message if running -join_io with decomposition other than graph decomposition

removed unneeded #include

changes need to deal with input files passing in "automatic" file type

the bridging scale code in the development module can now be enabled/disabled
in DevelopmentElementsConfig.h with the BRIDGING_ELEMENT_DEV flag, which is
disabled by default

dump one Metrowerks CW for each input file

profile TakeParameterList, as well as Solve, when running the MW profiler

revised prompts when generating spatial decomposition

Partition::kAtom replaced with PartitionT::kIndex

updated to reflect change in PartitionT.h

clear trailing new line from console input

reivsed handling of the decomposition method. The code now calculates a
spatial decomposition, though it cannot compute with it yet.

merging changes from BRANCH_david_1: revised handling of model decomposition
for parallel execution

clean-up before merge

cleanup before merge with main branch

clean up

(1) corrected task variables
(2) moved echo of partial geometry file

continued debuging parallel standard tahoe

revised initialization process so pointer to dynamically allocated objects
are stored before initialization so memory is properly cleared if an exception
occurs during initialization. Old implementation leaked memory under these
conditions.

added some debugging code

debugging the parallel run

Made adjustments such that benchmarks pass at level.0

added some notes

(1) revised handling of the different types of FEManagerT's
(2) removed bridging scale calculations from FEExecutionManagerT

enable quasistatic bridging scale capability

merging changes from BRANCH_parameter_4. Tahoe is now XML-powered.

revisions from the benchmark module

added comment to bridging code

removed unneeded code

revisions needed because a file path, not a stream, is passed to FEManagerT
on construction

set path to root file

removing unneeded methods

corrected file paths for parallel decomposition and joining

continued merge of changes from BRANCH_parameter_3

merging changes from BRANCH_parameter_3

changes to restore parallel capability. Decomposition working again.

1.  Merging BRANCH_LJ3D to main line of Tahoe

merging changes from BRANCH_penetratio_2: pass command line options when
constructing FEManagerT's

merging changes from BRANCH_penetration_2: changes need to make command
line options accessible from within the FEManagerT

merging changes from BRANCH_bridging_6: verify time integration parameters
are the same for both the atomistic and continuum solver for bridging scale
calculations.

1.  Clean up

1.  Modified files to account for ifstreamT.h and ofstreamT.h changes on main line

updates for changes to fstreamT/ifstreamT/ofstreamT

enabled automatic load control in MultiManagerT calculations

merging changes from BRANCH_bridging_5

1.  Changes to GhostOn and GhostOff maps

1.  Changes to accommodate 3D LJ bridging scale simulations

1.  Committing changes from BRANCH_BS3D to merge with mainline

merging changes from BRANCH_parameter_2. BRANCH_parameter_3 continues on
with revisions from the main line since BRANCH_parameter_2.

make command line options accessible through FEManagerT and ElementSupportT

merging changes from BRANCH_penetration_1

echo parameters to XML after parsing and before validation to aid with
debugging

revised dArrayT::operator=

1.  Changes to EAMT to allow multiple types, or interatomic potentials

revisions needed for changes to the toolbox

report description file stats

revised dArrayT::operator=

changes needed for revised dArrayT::operator=

changes needed to avoid ambiguity with dArrayT::operator=

1.  Final checkins before new branch

1.  Further deliniations of 3D EAM-specific stuff in FEExecutionManagerT

1.  Revised FEManagerT_bridging::SetExternalElecDensity such that pointers to embedding force and electron density are directly set to EAMT class pointers such that Assemble functions are not needed

1.  Wrote code such that FEManagerT_bridging atoms can access electron density/embedding terms calculated in FEManagerT_bridging continuum
2.  Wrote code such that FEManagerT_bridging atoms can write into EAMT::fElectronDensity and EAMT::fEmbeddingForce for ghost atoms

1.  Preliminary changes (primarily in FEManagerT_bridging) to allow continuum approximations of electron density and embedding terms for EAM calculations at ghost atoms

added missing header files

merging changes from BRANCH_bridging_4:
(1) corrected atomic volume in virial stress calculations and pair potential
    Cauchy-Born materials
(2) added flag to control whether image nodes are included in the list of
    non-ghost nodes
(3) corrected cubic spline window function with spherical support
(4) added selectable window functions and parameters for MeshfreeBridgingT

added flag to control whether image nodes are included in the list of
non-ghost nodes

added flag to indicate whether image nodes should be included in the list
of non-ghost nodes

1.  Write calculated THK GA displacements into field

revised construction patterns

1.  Added beginnings of 3D THK implementation

1. Further 2D/3D splitting.  Correct implementation of InterpolationNodes3D

1.  Wrote code to prevent "image atoms" due to periodic BC's from entering NonGhostNodes() array

changes to construct and initialize problem using XML input

1.  Further 2D/3D revisions

1.  Further modifications for 2D/3D bridging scale split

1.  Initial BRANCH_BS3D modifications

1. Check-ins to fix issues with previous updates

Changes to account for:
1.  Generalization of bridging scale code to 3D for arrays and matrices
2.  Usage of MD atomic masses in calculation of bridging scale projection matrix and global mass matrix

changes for switch to XML input

split communicator before running decomp and join

-j HEAD, altered slip vector and strain calculations for Periodic BC's

merging changes from BRANCH_bridging_3:
(1) added driver for solving coupled system monolithically using the
    MultiManagerT
(2) removed code for the output map which is now computed by each
    processor instead of being written/read from a file

The output map is now computed instead of written/read to/from a file

initialization of projection/interpolation data for the monolithic
solution scheme is handled within the MultiManagerT

added capability to run FE/MD coupled problems with a monolithic solution
scheme using the MultiManagerT to drive two sub-FEManagerT's

Fixed Strain calculations, added calcs to EAM. Still not compatible with benchmarks

a little test code

1.  Committing branch_thk_3 changes to merge with main tahoe branch

1.  Added #if and #endif statement to around debugging sections of code

Initial commit of changed files for BRANCH_cjkimme_1. This is a catch-all holder
for finished and unifinished implementation of atomistic conveyor belts for steady-state
crack propagation, prescribed motion read in from Paradyn output files, barostats
and configurational thermostats for ParticleT elements, and fine/coarse scale
decompositions of arbitrary dynamical variables.

1.  Final summer checkin

1.  More modifications for dissipation calculations

1.  Modified code to calculate dissipation due to THK forces

1.  Added code to calculate dissipation due to THK forces

write XML Schema file if -xsd option is passed or entered at the prompt

spit out .dtd file if tahoe is run with -dtd option or if -dtd is passed
on the input line

added -dtd command line option which causes tahoe to dump a DTD file of the
available input parameters

1.  checking in changes to allow different problem types

1.  bridging scale specific changes for the branch

fixes for building withouth the development module

revisions to fix builds without the development module

1.  Committing bridging scale changes from BRANCH_THK_2

1.  Further changes for multi dimensional bridging scale

1.  Changes to allow for multiple finite elements in each spatial direction for bridging scale calculations

1.  Added options to disable ghost nodes during bridging scale calculations

1.  Modifications

1.  Nightly commit

1.  Small changes

1.  Modified to call THKForce calculation function

1.  up to date

1.  Call gaussian wave function from FEManagerT_THK

initialize variable before using it in RunStaticBridging

1.  Modified to call new FEManagerT_THK::THKForce function

1.  Modified for combined predictor/corrector function call in FEManagerT_THK

1.  Added debugging comments

1.  Using SetFieldValues to write ghost atom displacement into field instead of SetKBCNodes

1.  Added predictor/corrector time integration for FEM solution at MD boundary atoms

partial changes for revised FEManagerT_THK::InterpolationNodes and new
FEManagerT_bridging::SetGhostNodeKBC.

1.  Adding changes to integrate interpolated FEM solution at MD boundary/ghost atoms

1.  More commits to test bugs

1.  Committing changes to check bugs in in put file

1.  Modified InterpolateField to be able to perform interpolation on arbitrary field

small bug fixes

1.  Updated DynamicBridging to function with new InitialCondition Tahoe implementation

added preprocessor directives to disable the THK code if the development
module is not installed.

merging changes from BRANCH_THK_1

explicit calls to ReadRestart in RunBridging not needed anymore because
it has been incorporated into InitialCondition

1.  Added function InitialProject to generate initial FEM displacement field from MD, and write into global displacement array

1.  Returned SetExternalForce to original design, generated nodes array to pass in
2.  Revised BridgingFields to return total u obtained from adding projected fem + fine scale

1.  Modified SetExternalForce to only pass in (field, dArray2DT&) because active node list should not change
2.  Implemented SetExternalForce in FEExecutionManagerT

1.  Modified Ntf function to accomodate global to local mapping for both nodes and atoms
2.  Added functinality to compute N^{T}f(u) multiplication in FEExecutionManagerT.cpp

Added Ntf function to calculate global interpolation matrix N_{I}(X_{\alpha}) for bridging scale calculations in FEManagerT_bridging

Finalize InternalForce for calculation of FEM internal force for bridging scale calculations

finished method to calculate and return the internal force contribution
to the residual

1.  Commit changes for synching with Patrick

1.  Added documentation for proposed bridging scale calculation flowchart

1.  Added functionality to compute total bridging scale solution u

1.  Some tahoe namespace fixes
2.  Some updated files relating to bridging scale

1.  Implemented RunDynamicBridging for basic coupled MD/FE capabilities

small revisions to split into static and dynamic bridging scale calculations

1.  Added split depending on static/dynamic problem

If Tahoe is launched as

    tahoe -thk

(case sensitive), the input file is passed to a FEManagerT_THK for execution.

fixed division by zero if the initial residual is zero.

corrections so that Tahoe builds with BRIDGING_ELEMENT disabled.

merging changes from BRANCH_bridging_1: added new "multi-Tahoe" execution
modes where two instances of the Tahoe simulation object are created to
run separate calculations that are coupled.

(1) corrected handling for double %% to identify bridging calculations
(2) corrected disabling of additional cross-coupling terms

(1) defined enum for different execution modes
(2) input file with leading %% is now recognized as a bridging scale
    simulation
(3) disabled all additional cross-coupling because it was incomplete

added more strongly coupled solution method

started implementation of cross-coupling matricies between the active
and follower displacements in the atomistic and continuum domains.

clean-up

small change to output the number of bridging cycles and the number of
iterations for the atomistic and continuum problem per load increment.

changes to allow SolveStep to be called multiple times for the same
time increment, so multiple Tahoe's can be solved in a staggered
fashion.

1. added separate log file for convergence of coupled problem.
2. rewrote solution scheme to be correctly staggered.

disable error check on return values in main bridging scale loop. Now
calculates solution though convergence of 2 sub-problems is slow.

All the bridging code is in place, but hasn't been run through yet.

tying atom and continuum solvers. Put a time sequence solve loop on the
execution manager. Just need that projection.

Minor fixes here and there to get the bridging simulation through the
initialization phase.

code to take a bridging scale calculation through the initialization phase

beginning implementation of prototype bridging scale calculations

passing new -decomp_method option overrides previous settings

merging changes and conflicts from BRANCH_atom_1

added command line option -verbose which triggers CommunicatorT do
verbose logging

minor tweaks to MP execution:
(1) will run decomp and join on rank 0 if run with size > 1
(2) skip comment lines in batch files

moving more work for parallel execution into CommManagerT

setting up CommManagerT to handle different decomposition schemes

decomposition type enum moved to PartitionT

corrected test for when to prompt user for decomp method with parallel
jobs. Mainly, don't want a prompt if the decomp files exists or particular
ones aren't needed (split vs. joined I/O)

don't prompt for decomposition method with parallel runs unless
decomposition is needed

less memory usage during Decompose_atom

(1) added atom decomposition method
(2) prompt for decomposition method or look for the method as an option
    -N , where N=0,1,2 following the flag -decomp_method
(3) write partition information after decomposition is calculated by
    FEManagerT_mpi is written. Info used to be written within FEManagerT_mpi.
(4) more efficient writing of partial geometry files. Previously, global
    model file would get scanned multiple times. Now a single model
    manager is initialized once and used repeatedly.

replaced flag __MPI__ with flag __TAHOE_MPI__ because the old flag is pre-defined in some environments

changes argument in AssembleRHS from dArrayT to nArrayT<double> so sub-classes
can be assembled as flat arrays

merge changes from BRANCH_paklein_1:
(1) update to use ExceptionT
(2) removed deprecated Allocate statements. Replaced with Dimension

replaced deprecated calls to Allocate with the new method Dimension in
classes from toolbox/abc

updated to ExceptionT

merging changes from BRANCH_B_1:
(1) moved calculation of B from shape functions to elements
(2) added initial version of Simo's Q1P0 element. No consistent tangent.

workaround for getline bug in GCC 3.1

toolbox header Constants.h has been renamed toolboxConstants.h

(1) incorporating CommunicatorT to handle MP messages
(2) restored exception handling for MP jobs

started replacing direct calls to the MPI library with the
CommunicatorT class (toolbox).

Merging BRANCH_name_1 onto main branch. The code compiles and
passes the benchmarks on SGI, linux, Sun, and CodeWarrior in Win2K.

Initial wrapping of namespace Tahoe around libtahoe. It currently compiles
on linux and passes the benchmarks. Other systems haven't been tried yet.

(1) the output set to processor map file, used during -join_io, is now
    associated with the parameter file name instead of the geometry file.
(2) corrected issue associated with setting the output time stamp when
    writing output data. Previously, the time was only set during WriteOutput.
    Now a default value is set during InitStep. The old was causing trouble
    with history node output between calls to WriteOutput.

Updated to agree with new ExodusT.

revised some things with command line option handling

(1) more generalized join
(2) better support for swapping command line options

all references to element/node/side/output sets have been updated to
use StringT id's instead of int's

removed all references to global model file (not needed anymore)

corrections for runtime joining

(1) changes to work with revised JoinOutputT
(2) even with -split_io, tahoe will try to create partial geometry files if they are missing

no longer writing global model file. Not needed to join files with the revised OutputSetT and ModelManagerT

still need global model file for now.

parallel fixes with new I/O

changes to work with revised OutputSetT, which now hold information about the element blocks in each output set. Also, the number of element blocks in the output is no longer the number of output sets.

(1) command line options (anything starting with '-') are now accepted at the prompt. Additionally, "-run", "-decomp", and "-join" are treated at mutually exclusive.
(2) METIS used by default if installed unless "-no_metis" is passed in. No more user prompt for method. This was causing unexpected prompt during parallel execution.

merged BRANCH_model_1 and resolved conflicts

1. Modify ReadParameters and WriteParameters
2. Modify Initialize
3. Remove SetInput
4. Move fInputFormat and fModelFile to FEManagerT_mpi, could eventually
remove entirely when decomposition uses ModelManagerT.
5. Remove functions to return fInputFormat and fModelFile.

fixed decomp for MPI

pass method flag for decomposition functions

fix to handle input files with no extension

Output statistics of decomposition

Added clear to end of line after prompting for number of partitions to join.
Trailing newline was flushing input at next prompt.

Added fix to clear newline after prompting for size in RunDecomp_serial. Otherwise, next read from cin is empty.

initial check-in

Initial revision