Adding a class that will encapsulate the discretization methods.
696 lines of code changed in 2 files:
Reverting to previous version of DQMOM class until I figure out what's wrong with new version
59 lines of code changed in 2 files:
user can now input physical coordinates as opposed to cell coordinates
0 lines of code changed in 2 files:
Also updating LU solver for DQMOM
28 lines of code changed in 2 files:
Adding calculation of norm of the residual to DQMOM solver.
127 lines of code changed in 2 files:
added option to input physical coordinates for the starting and ending pts.
I'm not completely happy with this. Computing the cell indicies from the input
points occurs _before_ any bullet proofing takes place. The bullet proofing
(do level exists for all time steps) should be placed immediately after parsing the
arguments.
37 lines of code changed in 1 file:
Initial import of the DQMOM solvers into Arches along with a few model terms.
6285 lines of code changed in 49 files:
Adding the LU solver.
282 lines of code changed in 2 files:
Fixed particle data memory leaks
107 lines of code changed in 4 files:
Added info about the combination of libraries.
0 lines of code changed in 4 files:
Rearrange order of variable initialization to quiet compiler warning.
0 lines of code changed in 2 files:
updated ups files to pass the ups_spec
- changed path to sus.
6 lines of code changed in 6 files:
added newline to test svn commit
0 lines of code changed in 2 files:
Added the ability to turn on/off tile growth.
91 lines of code changed in 2 files:
Added the ability to specify the collection of particles with the <collectParticles> tag. The patchFactorParticles algorithm is now obsolete but is included for backwards compatability. The same algorithm can now be achieved by using patchFactor with collectParticles turned on.
7 lines of code changed in 1 file:
Factored out the multiply by the identity matrix. This saves a few floating point operations. This should not change the answers.
2 lines of code changed in 1 file:
preprocess_Lodi_BCs()
getICEMaterials -> getMaterials
The user is inputting the material index not the index for just the ICE matls.
Confusing enough?
1 lines of code changed in 1 file:
Added partial code for preventing memory leak
413 lines of code changed in 6 files:
Adding another conditional to check for division by zero
22 lines of code changed in 1 file:
Lowered the small number so that it is 5*machine precision. Eliminated a floating point division from the termination condition. Eliminated the iteration count and the corresponding termination condition. Terminate if the convergence has stopped.
0 lines of code changed in 2 files:
eliminate excess chatter
0 lines of code changed in 2 files:
octave script that computes the exact solution to the Couette-Poiseuille flow problem
and compares it to simulation data
201 lines of code changed in 1 file:
Fix typo in the outputProblemSpec
1 lines of code changed in 1 file:
Instead of just falling into the "weibull" option if nothing else fits,
use an "else if" to make sure that was what was intended.
1 lines of code changed in 1 file:
Make alpha an optional input with default 0. I don't think I've ever
not given it a value of 0 in the input file.
1 lines of code changed in 1 file:
fixed a parsing bug for the domain length. You can't delete the letter 'e' from
values that have exponents in them.
0 lines of code changed in 2 files:
Variable name change that should have been committed a long time ago.
centerOfMass -> TotalMomentum
1 lines of code changed in 1 file:
problemSetup
- added bulletproofing so the user can't apply LODI bcs to "all" materials
addRequires()
- require gammaLabel in CC_exchange portion
preprocess_Lodi_BCs()
- corrected how we determine the material index. This bug surfaced when we switch the
location of the ICE and MPM material sections in the ups file.
8 lines of code changed in 1 file:
raised the convergence criteria to 10x machine precision.
1 lines of code changed in 1 file:
increased the convergence criteria.
1 lines of code changed in 1 file:
Fixed a couple bugs in the cubeRoot function.
Wrote a CurbeRootTest program which compares the time and accuracy to pow.
139 lines of code changed in 4 files:
Update the code to reflect the changes that Cherie made to the input files
yesterday. d_adiabaticHeating is no longer a supported option. If one wishes
to turn off conduction, use
<DoExplicitHeatConduction>false</doExplicitHeatCondcution>
2 lines of code changed in 4 files:
Some changes for periodic NC datasets
29 lines of code changed in 1 file:
Added changes for periodic and no extraCell datasets
582 lines of code changed in 3 files:
Quiet compiler warning... #if 0 code that doesn't do anything.
2 lines of code changed in 1 file:
Fixed compiler warning (put unsued var inside #if). Cosmetics... white space, coding standard...
24 lines of code changed in 1 file:
Commit of configure that corresponds to new configure.ac
1 lines of code changed in 2 files:
Cleaning up more of the Uintah tree.
Don't need/use ogl stuff anymore. Don't need the osx/ directory as
the hack isn't necessary anymore (We are 2 major OS versions past it).
1 lines of code changed in 6 files:
Fixing the driver so that it uses a set of moments that works OK
0 lines of code changed in 2 files:
Updating DQMOM matlab routines
185 lines of code changed in 5 files:
Fixing some indexing problems
15 lines of code changed in 1 file:
Fixing a model source term bug - don't accumulate model terms...
2 lines of code changed in 1 file:
Combined the two cubeRoot functions into one function with a default parameter. This eliminates some code duplication.
11 lines of code changed in 3 files:
Undid last patch and implemented it a better way.
2 lines of code changed in 1 file:
Don't output the AMR overhead on non-AMR problems.
5 lines of code changed in 1 file:
Getting rid of NormalForceBC from PhysicalBCs. I wrote it, I can't figure
out what I was thinking, so I can't document it, it isn't used. Outta here!
2 lines of code changed in 6 files:
Added new cubeRoot functions for quickly computing the cubedRoot of a double.
87 lines of code changed in 2 files:
Removed >= <= operators in IntVector.h.
Changed the order of foreign Grid variable link list in DW.
39 lines of code changed in 4 files:
Fixing a negative sign; also adding print statements for A, B, and X matrices for debugging purposes
27 lines of code changed in 1 file:
Adding a method to get the quadrature node associated with a DQMOM equation
0 lines of code changed in 4 files:
Adding LU decomposition solver (Crout.m) to compare Crout's method to Matlab's A^-1*B solver
156 lines of code changed in 3 files:
Remove MALLOC_TRACE_TAG committed earlier.
0 lines of code changed in 2 files:
More robust checking if we are using petsc version 2.3.
18 lines of code changed in 2 files:
Use the MallocTrace library.
70 lines of code changed in 2 files:
M Core/Grid/Variables/Variable.h
M Core/Grid/Variables/Variable.cc
M Core/Geometry/IntVector.h
M CCA/Components/Schedulers/OnDemandDataWarehouse.cc
Add a next pointer for vars in DataWarehouse.
This linked list of vars is to store foreign Grid Variables received from same patch but with different range.
We need multiple ranges(windows) of data under same label for reordering task executions and MPI receives.
M CCA/Components/Schedulers/MPIScheduler.cc
M StandAlone/inputs/UPS_SPEC/ups_spec.xml
MPI queued task scheduler works now and can be turned on by tag <useExternalQueue_>.
Default is still the old fix-order scheduler.
117 lines of code changed in 5 files:
Quiet debug spamming...
0 lines of code changed in 1 file:
Merged LICENSE and LICENSE_CSAFE into one file (LICENSE).
0 lines of code changed in 3 files:
1) fixing bug in LU solver related to vector access
2) adding check in DQMOM problem setup for moment index size and number of indices
3) from SpatialOps.cc, removing initialization of weights/weighted abscissas in half the domain to be 0 (so we don't get a bunch of messages about A being singular)
38 lines of code changed in 3 files:
No longer passing in the varLabel into the transport equation. This should simplify things a bit.
23 lines of code changed in 14 files:
Fixing a bug in DQMOM class related to order of construction of A and B in AX=B solver
20 lines of code changed in 1 file:
- Fixing some major problems with the AX=B construction process
- Adding ability to incorporate phase-space diffusion in B matrix
121 lines of code changed in 2 files:
Fixing bug in source term calculation
1 lines of code changed in 1 file:
Adding "DQMOM sandbox" matlab script to SpatialOps
337 lines of code changed in 2 files:
Dd: On Macs, it isn't '.so', it is '.dylib' (so use generic var created for this).
1 lines of code changed in 1 file:
Removed Tabs (Dd)
4 lines of code changed in 1 file:
Fixed issues with multi-level NC data. Some memory clean up also done.
600 lines of code changed in 3 files:
Fix max link.
0 lines of code changed in 2 files:
contents moved to doc/ice.tex
0 lines of code changed in 1 file:
Fixing major sign error
1 lines of code changed in 1 file:
1) Fixed bugs in Ax=b solver.
2) Fixed sign error in DQMOMEqn.cc
3) Fixed src*vol in DQMOMEqn.cc
56 lines of code changed in 6 files:
Adding the missing weights assignment
5 lines of code changed in 1 file:
Source term wasn't being retrieved properly for higher order time stepping.
Also, there is missing code in DQMOM.cc. The source term for the weight equations are not being filled in.
50 lines of code changed in 7 files:
1) Fixed a bug in the multi-step time integrator. Now using old RK values rather than old_dw values in the intermediate steps.
2) Now passing in the ExplicitTimeIntegrator into the builder and then into the actual equation class rather than trying to set it afterwards.
283 lines of code changed in 10 files:
M CoalModels/PartVel.cc
M CoalModels/PartVel.h
- Fixed some bugs in the Particle velocity model. The model is adding to the magnitude of velocity...do we want to try adding to the components?
M TransportEqns/ScalarEqn.h
- Passing the timesubStep into the task properly.
M TransportEqns/EqnBase.cc
M TransportEqns/EqnBase.h
- Added a mechanism for storing initial conditions (still need to allow for reading in from the input file).
- Added an "old" transport varLabel for the RK time stepping.
M TransportEqns/DQMOMEqn.cc
M TransportEqns/DQMOMEqn.h
- Added upwind and super_bee convection schemes.
- Passing the timesubstep properly now.
- Fixed a sign error in the convection scheme addition to the RHS.
- Perform data copy after a RK step.
M TransportEqns/ScalarEqn.cc
- Passing the timesubstep properly now.
- Added an "if" statement to the sched_initialize (only perform when timesubstep = 0)
- Note: Still need to add flux limiters here.
M ExplicitTimeInt.h
- Now passing alpha and beta into the time averaging method. I was having some weirdness here that should be revisited (potential memory bug?).
M SpatialOps.cc
- Now setting the time integrator for the DQMOM equations. Don't ask me how this actually worked before :p
- Initializing particle velocities in actuallyInitialize.
- Right now there is a bunch of problem specific initialization going on that we will want to generalize in the future.
- Computing stable time step using a convective CFL. Will want to refine this (perhaps a different norm? Add diffusive CFL) soon.
M SourceTerms/SourceTermBase.h
- Passing timesubstep properly now and storing it in the base class.
M SourceTerms/ConstSrcTerm.cc
M SourceTerms/ConstSrcTerm.h
- Adusting to changes made in the base class.
M StandAlone/inputs/UPS_SPEC/spatialops_spec.xml
- Added more spec for the particle velocity model.
439 lines of code changed in 18 files:
Fix to my bad merge from previous version.
0 lines of code changed in 2 files:
Adding ability to grab moments from input file for DQMOM; also updating SpatialOps input file to include particle velocity and moment specifications.
9 lines of code changed in 1 file:
M CCA/Components/OnTheFlyAnalysis/particleExtract.cc
- Fixed compiler warnings. Use "()" around portions of if statements
to make them more clear (make explicit order of evaluation).
- Added 'default' to switch statement to quiet compiler complaint that some
options were not handled.
M CCA/Components/ICE/CustomBCs/LODI2.cc
- Fixed compiler warnings. Use "()" around portions of if statements
to make them more clear (make explicit order of evaluation).
M CCA/Components/Arches/Properties.cc
- Indented and added "{}" around some if statements. Quiets compiler
warnings and makes code easier to follow.
M CCA/Components/Arches/MCRT/ArchesRMCRT/RMCRTRadiationModel.h
- Added "{}" around some if statements. Quiets compiler warnings and
makes code easier to follow.
M Core/Grid/Variables/constVariable.h
M Core/Grid/Variables/constVariableBase.h
- Make getBaseRep() const. This allows for making a vector of
constVariable pointers like this:
constCCVariable<int> testCC;
vector<const constCCVariable<int>*> vCCVars;
// Get the testCC variable.
new_dw->get( testCC, lb->partCountLabel, 0, patch, Ghost::None, 0 );
// Create an empty constCCVariable.
constCCVariable<int> * testCCp = scinew constCCVariable<int>();
// Copy the pointer to the data (testCC) into our new variable.
testCCp->copyPointer( testCC );
// Save the new variable in a vector for future use.
vCCVars.push_back( testCCp );
- Commented out duplicate(ish) copyPointer() that can't be used anyway
because the compiler complains that it is ambiguous with the other
version of copyPointer(). (The parameters Variable& and
VariableBase& can't be distinguished.)
116 lines of code changed in 8 files:
Adding a more robust LU decomposition solver/routine; also, a little cleaup of the DQMOM class.
176 lines of code changed in 4 files:
Without SCIRun, we don't need on-the-fly-libs anymore...
0 lines of code changed in 3 files:
Only fail test 6 if datafiles are gzip-compressed.
0 lines of code changed in 2 files:
Need teem (and png) lib to link.
53 lines of code changed in 13 files:
Don't use Packages anymore.
2 lines of code changed in 1 file:
Still need Dynamic compilation stuff (until Fields go away (shortly)). Putting it back.
2472 lines of code changed in 5 files:
D DynamicLoader.cc
D DynamicLoader.h
M sub.mk
D DynamicCompilation.cc
D DynamicCompilation.h
- New 'Uintah' does not use dynamic compilation... so removing associated files.
M Environment.cc
- Don't use ITCL, LOAD_PACKAGE, THIRDPARTY anymore.
0 lines of code changed in 7 files:
Removed references to SCIRun 'Packages' as they don't exist anymore.
M configure.ac
M configure
- Require autoconf 2.63 to maintain consistency on the 'configure' file.
- Add "--with-teem" arg specification. (Removes configure complaint when using --with-teem.)
- AC_SUBST( name ) (in autoconf 2.63?) puts 'name ' (space at end) so use (name).
- NOTE: turned off svn property svn:eol-style because it was causing svn to have
a conniption fit, due to, I believe, the fact that autoconf 2.63 adds a ^M
into configure (so it can test for the end of line character). Unfortunately,
svn isn't smart enough to handle this correctly, so it just gives errors... sigh.
M aclocal.m4
M include/sci_defs/environment_testdefs.h.in
240 lines of code changed in 5 files:
Fixing a bug and adding a moment that won't confuse the LU solver
15 lines of code changed in 1 file:
All these files are empty... I think they were consolidated in to GridVariableBase, but not deleted...
0 lines of code changed in 10 files:
M ScalarSolver.cc
- Fixed compiler warnings:
-- Commented out unused variables.
-- Use () around ?: notation... Also I assume this code needs 'boolean
or' ("||") and not 'binary (bitwise) or' ("|")... Please verify.
5 lines of code changed in 1 file:
using package setspace instead of doublespace
0 lines of code changed in 17 files:
Divide the weighted abscissa with the weight to get the right length.
38 lines of code changed in 1 file:
Modifying the way DQMOM.cc stores weights and weighted abscissas
72 lines of code changed in 1 file:
Fixes for node centered data
387 lines of code changed in 3 files: