Use less expensive string instead of ostringstream for message buffer inside a computation loop
7 lines of code changed in 1 file:
Added ability to rotate the axis of the torus about one other axis. Still not
totally general, but as close as I need it for now.
26 lines of code changed in 3 files:
Remove the extra Tag:: in the scope resolution.
2 lines of code changed in 1 file:
Fix FakePetscSolver so we can compile without Petsc
1 lines of code changed in 1 file:
Allow user to choose the axis around which the donut sits. Still requires
alignment with the x, y or z axes.
47 lines of code changed in 2 files:
modified some programming deficiencies in simplifiedGeoModel.cc
36 lines of code changed in 1 file:
Added swirl boundary condition.
Started documentation of boundary conditions in Arches.
136 lines of code changed in 2 files:
added and cleaned up the logic to execute
matrixCreate() and destroyMatrix()
at the appropriate times.
This will fix the highwater memory usage RT errors.
All ARCHES-dbg tests passed successfully!
20 lines of code changed in 2 files:
switcher.cc
- attach modelMaker port to turn on the burn models.
ex3_phases1_3.ups
- saving modelMass_src, totalMassBurned, totalHeatReleased.
4 lines of code changed in 1 file:
Adding a torus geometry piece. At this point, the major axis of the torus
is aligned with the z-axis. Will generalize as the need arises.
311 lines of code changed in 6 files:
remove a cout statement.
0 lines of code changed in 1 file:
cleanup
2 lines of code changed in 1 file:
Change the location where we set the IO boolean for a field. This works with Devin's memory management magic.
27 lines of code changed in 1 file:
MPMARCHES:scheduleMomExchange()
- only compute d_cellInfoLabel for the arches matl.
PressureSolverV2.cc
- initialize d_indx = -9
- addHydrostaticTermtoPressure()
use locally defined indx instead of d_indx. This task is called by
MPMARCHES and d_indx isn't defined.
UnknownVariable.cc
- removed conditional around adding matl index to exception message.
16 lines of code changed in 4 files:
The pressure solver has been segregated so the major sections of the solve are
now individual tasks. The tasks are:
sched_buildLinearMatrix
sched_setRHS_X_wrap
sched_SolveSystem
schedExtract_X
sched_normalizePress
sched_addHydrostaticTermtoPressure
In the previous implementation these sections were place inside of a loop
over patches. This change should not change any of the answers.
To reverse this change execute:
svn merge -r47642:47641 https://gforge.sci.utah.edu/svn/uintah/trunk/src/CCA/Components/Arches
86 lines of code changed in 16 files:
Update copyright date.
128 lines of code changed in 243 files:
added bulletproofing for uninitialized variables. This doesn't work for bools
and isn't perfect.
88 lines of code changed in 1 file:
added missing variables (time_multiplier & wmmsLnError) from previous commit
22 lines of code changed in 1 file:
Added three MMS problems as runtime options in MPM tree.
1018 lines of code changed in 20 files:
columnate
alphabetize
343 lines of code changed in 1 file:
Shelving Julien's commit until we can resolve the differences in the RT
or determine if we need new GS's and tests.
129 lines of code changed in 15 files:
Fixed C++11 variable initialization problem
0 lines of code changed in 4 files:
Updated FakePetscSolver to have a new dummy print() function.
5 lines of code changed in 1 file:
Implementing a better formualtion for weighted DQMOM
370 lines of code changed in 15 files:
* Fixed bug with restart. Objects were being allocated in paramInit but needed to be done elsewhere.
* Added the inverted tag for solids
* Added a simply wall type.
* velocity is now all specified inside of the velocity generator for all components.
* fixed some compiler warnings
155 lines of code changed in 5 files:
Add a list of fields that are saved by the uintah data archiver. Use that list to specify which fields in the tree are persistent. This in turn will be used in the memory management work being done on ExprLib.
45 lines of code changed in 6 files:
Pruned and sorted #includes.
12 lines of code changed in 1 file:
Unintialized intrusion object causing issues. Now using a boolean to
indicate on/off. Maybe find a more elegant solution later?
Also, memory cleanup needed.
94 lines of code changed in 8 files:
Initial commit of the new intrusion boundary mechanism. This new
version allows for more sophisticated boundary conditions on specified
faces of the intrusions. Currently, only flat profiles of scalars and
momentum are implemented.
1040 lines of code changed in 10 files:
Add sin(t) expression for use in verification of time integrators.
135 lines of code changed in 2 files:
Getting the right guess for X.
-added more printTasks and printSchedule
The only RT tests that fails is periodic.
82 lines of code changed in 2 files:
added another instance of printTask()
0 lines of code changed in 4 files:
modified simplified geomodel is committed
55 lines of code changed in 1 file:
added print() that dumps out A, X and B for either Petsc or Hypre solvers. The file
file names include the integration phase name, iteration and timestep.
Currently turned off in PressureSolver.cc
60 lines of code changed in 7 files:
fix precision issues when reading UDA data into VisIt
25 lines of code changed in 4 files:
add INSTALL_TYPE switch to udaReader makefile generator, can be -public to install plugin with VisIt installation or -private (default) to install in /uufs/chpc.utah.edu/common/home/u0033785/.visit
5 lines of code changed in 1 file:
More work on the rewrite of PressureSolver.cc
Status: All RT tests run to completion and the pass/fail score is shown below:
pass: constantMMS"
pass: almgrenMMS"
fail: periodic"
fail: helium_RT"
pass: methane_explicit_table"
pass: methane8patch"
pass: methanePetscRadSolver"
pass: dqmom_test_1"
pass: dqmom_test_2"
pass: dqmom_test_3"
pass: dqmom_test_4"
pass: dqmom_test_5"
pass: methane_jet"
pass:acidbase_jet"
116 lines of code changed in 2 files:
Missing dependency in sub.mk.
1 lines of code changed in 1 file:
Add back in hypre includes.
4 lines of code changed in 1 file:
cleaned out #includes
0 lines of code changed in 1 file:
Add more useful information to error diagnostics in boundary conditions
4 lines of code changed in 1 file:
added bulletproofing.
- The user must specify 1 and only 1 side boundary condition.
0 lines of code changed in 4 files:
Renaming some variables.
4 lines of code changed in 2 files:
Calling the Glacier rqpart fortran code now in the correct place for
particle radiation properties.
42 lines of code changed in 8 files:
added input file extension .ups
1 lines of code changed in 1 file:
Variable Density:
The variable density has been added in this commit for the scalar transport equation and momentum equation. So, for solving these types of equation we will have to options of constant or variable density from now on.
wasatch_sepc.xml:
There are some important changes in this xml file in TransportEquation block which may cause your input files to an error. You may need to review those changes and modify your existing input files and write your new input file with respect to them.
DensityCalculator Expression:
This expression has been added in order to calculate density from that thermodynamic tables (from TabProps) having the solution variable which is (density * primitive variable), specially for solving transport equations in variable density cases.
Multiplier Expression:
This expression has been added to calculate the product of to other tags (expressions). This is specially useful for calculating the initial condition of the solution variable from the initial condition of primitive variable and the density which can be a case in variable density problems.
DiffusiveVelocity Expression:
This expression calculates \f$ \Gamma_\phi * \nabla \phi \f$ instead of diffusive flux which is \f$ \rho * \Gamma_\phi * \nabla \phi \f$. In scalar transport equations when we have constant density or when we want to solve weak forms of the equations which will be implemented in future, or in energy transport equation for temperature we don't have the accrual diffusive flux term. The term that we need in these cases is actually diffusive velocity which excludes the density.
1980 lines of code changed in 27 files:
Fixes for Mac OSX Lion... it is the same as SNOW LEOPARD, but because
the os version number was incremented, we need to check for that.
(Note, a fix was at one point committed for this in 'configure' but
(I think) didn't make it into configure.ac. Now it is in both places.)
M configure.ac
M configure
- For John: Added in (c) 2011. :)
- Add to SNOW_LEOPARD_OR_LATER the check for Lion.
M include/sci_defs/osx_testdefs.h.in
M Core/Grid/Variables/ReductionVariable_special.cc
M Core/Disclosure/TypeUtils.cc
M Core/Util/Endian.cc
M StandAlone/tools/puda/varsummary.cc
- Use SNOW_LEOPARD_OR_LATER flag...
28 lines of code changed in 12 files:
This file should have been included in my previous commit.
- removed addToTimestepXML() and readFromTimestepXML() this is just duplication
of outputProblemSpec() and problemSetup()
0 lines of code changed in 1 file:
global:
- removed addToTimestepXML() and readFromTimestepXML() this is just duplication
of outputProblemSpec() and problemSetup()
Switcher (AKA Mexican Jail)
- final cleanup
Added comments
Changed some exception types from problemSetup -> InernalError
more descriptive exception messages
* I'm now playing my get out of jail free card.
58 lines of code changed in 7 files:
added switcher example 3 to the nightly and local RT.
changed filename so the RT test name would be more descriptive.
5 lines of code changed in 1 file:
Fixing bugs:
ClassicTableInterface -- extra (spurious?) line in the boundary
assignment.
Discretization -- psi where r should have been
Input files -- missing specification of x- face.
-- Julien and Jeremy
2 lines of code changed in 3 files:
more diagnostic output
added simpleBurn to the switch criteria in ups_spec.
8 lines of code changed in 4 files:
iterate()
- changed the convergence criteria. If the max values for
dispIncNorm or dispIncQNorm are less than their respective
tolerances then the system is converged. This fixes the issue
if dispIncNorm & dispIncNormMax are both epsilon.
0 lines of code changed in 2 files:
actuallyInitialize()
- added a warning that anything other than linear interpolation may not work.
- added bulletproofing. Throw an exception if the proper number of ghost cells
are not specified for a particular interpolator.
33 lines of code changed in 1 file:
moved switcher related ups files to a new subdirectory (Switcher).
UCF.py:
fallout from the move.
8 lines of code changed in 1 file:
Fix a potential problem with a method in TaskInterface.
4 lines of code changed in 1 file:
problemSetup()
-turned scrubbing off for init variables.
initNewVars(()
- when initializing new particle variables get the particle subset from
the old_dw instead of the new_dw.
Global:
- more feeback to the user when component switching takes place.
30 lines of code changed in 1 file:
removeFixedDOFHeat()
added debugging code to output A & B to files
0 lines of code changed in 2 files:
Do not set the PETSC_ARCH unless the user specifies it, fixes problems on ember/updraft.
80 lines of code changed in 2 files:
Find petsc libs/include when not specified --with-petsc
9 lines of code changed in 2 files:
Fix test for Petsc SINGLE_LIB flag.
4 lines of code changed in 2 files:
Fix test for Petsc SINGLE_LIB flag.
4 lines of code changed in 2 files:
Add support for petsc version 3.
504 lines of code changed in 2 files:
Add the necessary tools to set the bcs for the momentum solver. Still need to check whether we need bcs on the momentum rhs or on the convective and diffusive fluxes separately. Also, add ability to solve under inviscid flow conditions by avoiding the constructiong of viscous stress expressions. This will help in debugging the momentum solver with boundary conditions.
767 lines of code changed in 8 files:
Allow user to specify output precision in lineextract and timeextract tools. Also, set default precision in timeextract to 16 to be consistent with that in lineextract. Finally, fix typos and formatting for the help section.
0 lines of code changed in 4 files:
Remove the petsc-uni and mpi-uni checks, since they are no longer valid with Uintah.
29 lines of code changed in 7 files:
cleaned out #ifs around old code.
removed unused variables and couts
0 lines of code changed in 2 files:
added bullet proofing. The user can no longer specify that the file is
binary and then try to read in a text file and spend X hours trying to debug
why that doesn't work!
removed all of the d_presplit code. All of the files are presplit.
Added comments and cleanup
0 lines of code changed in 4 files:
replaced try-catch statements with if-else.
*This should not change the answers
0 lines of code changed in 2 files:
local test == nightly tests
0 lines of code changed in 2 files:
MPMICE::actuallyInitialize()
- clarified exception message.
- #if 0 code that may go in the dumpster soon.
MPM::intializeCCVariables()
- cleaned out code that was never used.
- removed try-catch statements.
- cleaned up indentation
- added comments
* The answers should not change --Jim & Todd
95 lines of code changed in 5 files:
Need #include<cmath> for std::isnan. -Dav
1 lines of code changed in 1 file:
The local tests are the same as tne nightly tests.
removed all darwin tests
0 lines of code changed in 2 files:
Adding an operation to remove warning. Will settle this issue later.
2 lines of code changed in 1 file:
Compute the stripping factor with the conserved mixture fraction.
4 lines of code changed in 1 file:
Use d_initialData.rhoS to fix compiler error.
0 lines of code changed in 2 files:
Now compute alpha0 from the specified material density and the density of
the fully densified material, rhoS. Also fix a discontinuity in the
unloading pressure.
20 lines of code changed in 1 file:
Fix for Jeremy's new CoalModel fortran linking problem. Jeremy, please
see comments below:
M CCA/Components/Arches/CoalModels/sub.mk
In order for the fortran files .h (in this case: rqpart_fort.h) to be
created, another files needs to depend on it. Normally, there would
be a "#include <.../rqpart_fort.h>" line in one of the .cc files. I
didn't see this, so I assume you aren't actually using it yet
(probably waiting for the linking issue to be resolved. :) Because of
this, I arbitrarily made DragModel.o depend on rqpart_fort.h. By
setting this dependency, the make system knows to generate the
rqpart_fort.h file (from the rqpart.fspec file). Once you know which
file really depends on it (the one that #includes it), you should
update this sub.mk file accordingly.
M CCA/Components/Arches/CoalModels/fortran/sub.mk
The CoalModel/fortran directory is NOT being used as a separate
library (.so file). Thus we do NOT need the "include
smallso_prologue.mk" and "include smallso_epilogue.mk" lines (or the
PSELIBS or LIBS being set).
I also fixed the .h dependency. The file 'rqpart.h' does not exist,
and is not generated... however, 'rqpart_fort.h' is automatically
generated from the fspec file so that is the one we need here.
D CCA/Components/Arches/CoalModels/fortran/rqpart.f
A + CCA/Components/Arches/CoalModels/fortran/rqpart.F
I renamed the fortran from from .f to .F because our make rule is
based on '.F' files. I briefly looked at the build rules, and there
is one for '.f' files, but I think it is old and out of date. With
more time, it probably should be updated or scrapped.
M CCA/Components/Arches/sub.mk
Added in the CoalModel/fortran directory. Removed some tabs and alphabetized.
90 lines of code changed in 5 files:
fixed spacing
101 lines of code changed in 1 file:
binary table search implemented
106 lines of code changed in 1 file:
createPoint()
- Throw an exception if a point is being created outside of the computational domain.
24 lines of code changed in 1 file:
- fixed compiler warnings
- Throw an assert if you try to access a pixel/voxel outside of the image
- changed <= to < when looping over the image
15 lines of code changed in 1 file:
Revert checkins 47557 & 47558.
184 lines of code changed in 44 files:
The Hypre Setup phase is done at the beginning of the solve and
periodically depending on the value of <setupFrequency>.
665 lines of code changed in 28 files:
Add hypre changes.
8 lines of code changed in 1 file:
replace isnan with std::isnan to fix osx build
2 lines of code changed in 1 file:
- using proc0cout instead of cerr.
- more descriptive message to user.
- cleaned up indentation by gcd
25 lines of code changed in 1 file:
Added a sub.mk file for the new fortran stuff.
46 lines of code changed in 1 file:
Adding particle adsorbtion coefficient code. Not yet plugged into the
coal model.
624 lines of code changed in 8 files:
Typo in the denominator of the stripping factor.
1 lines of code changed in 1 file:
Fairly sure these cout statements in the constructor are leftovers from
a debugging session.
0 lines of code changed in 2 files:
User can now read in an auxilary file that contains two columns
intensity, scalar value.
Usage: ../tools/pfs/rawToUniqueGrains [options] <ups file>
options:
-b, -binary: binary output
-l, -littleEndian: input file contains little endian bytes [default]
-B, -bigEndian: input file contains big endian bytes
-auxScalarFile: name of file that contains two columns of data (intensity scalar)
# case: grains
# Number of blobs 75
# max diameter (cm) 0.0365792455209
# min diameter (cm) 0.00202786936452
# average diameter (cm) 0.00990896874237
# color equivalent spherical diameter (cm)
1.0 0.0365792455209
2.0 0.0031813408649
3.0 0.0113385664577
The scalar value is appended to all points with that intensity. The output files
now contain x, y, z, scalar.
It runs for a simple test problem but needs to be cleaned up and possibly templated
so the scalar can be a double, vector or whatever.
148 lines of code changed in 1 file:
added formatting to the DebugStream
10 lines of code changed in 1 file:
Removing some dead wood. Thanks to Todd for catching this as an issue in AMR.
0 lines of code changed in 1 file:
-Constructor
- addded bulletproofing, you can't run rmpmice and mpm
Global:
- added scheduleIntNewVars()
- For init variables switched from a bunch of vector<vector<X>>
to a map and a struct.
Still testing and thrashing around.
119 lines of code changed in 2 files:
Move timetask scheduling before rhs scheduling. This affects equations with explicit time dependence on the rhs.
19 lines of code changed in 1 file: