Fix compiler warning.
1 lines of code changed in 1 file:
Add an example of using surface grid iterators which mimics poisson1 version.
568 lines of code changed in 6 files:
M Arches/DQMOM.cc
Fixing bug in debug print statements
M Arches/Directives.h
Adding debug print statement compiler directive
12 lines of code changed in 2 files:
Copy d_curFace in the copy constructor.
0 lines of code changed in 2 files:
Additional funcitonality for the insertParticles task, including making it a
run time option and reading specification from a file.
62 lines of code changed in 7 files:
Updates to the WAIT_FOR_DEBUGGER code
0 lines of code changed in 6 files:
Some bug fixes
24 lines of code changed in 1 file:
Added a const.
0 lines of code changed in 2 files:
Supressing error message printing
2 lines of code changed in 1 file:
Add task to insert particles. Not quite finished. Scheduling of this task will
need to be an input file option. Currently scheduling is commented out.
100 lines of code changed in 2 files:
Comment out extraneous output line.
1 lines of code changed in 1 file:
Initial commit for global min/max display. Seems to work, but hasn't been extensively tested yet. Also not tested on multi-level UDAs.
190 lines of code changed in 3 files:
Fixing a memory leak
M CCA/Components/Arches/DQMOM.cc
Cleaning up, deleting some variables that weren't deleted before
M Core/Math/sci_lapack.cc
Changing some "delete"s to "delete[]"s
M Core/Datatypes/DenseMatrix.cc
Instead of always returning true, now returning INFO (Lapack variable indicating the solution status)
31 lines of code changed in 4 files:
gimp->GIMP
- fixed initial particle creation.
- Now computing the new lp in computeStressFromVelocity
- mms initialization
- set displacement, particle velocity, deformation gradient.
- Plotting diagnostic variables
- added MMS_velocity() function
******* HARD CODED TIMESTEP TO 1E-20 *********
The results for the MMS problem are wrong after the first timestep.
144 lines of code changed in 1 file:
Added patch/ files for improving particle rendering
1411 lines of code changed in 4 files:
Making a tag not required and adding an explanation of it
94 lines of code changed in 1 file:
Fix: Change p.failed -> p.localized to be consistent with previous change
2 lines of code changed in 1 file:
Change "pFailed_Label" to "pLocalized_Label" so that the naming and meaning is
more consistent with the ElasticPlastic model.
69 lines of code changed in 5 files:
M CCA/Components/Arches/CoalModels/SimpleHeatTransfer.cc
Fixing compiler warning
M CCA/Components/Arches/TransportEqns/EqnBase.h
Minor edits
M CCA/Components/Arches/TransportEqns/DQMOMEqn.cc
Bulletproofing problemSetup()
M CCA/Components/Arches/DQMOM.cc
Fixing bugs in moment calculation
Rearranging get() calls so they aren't happening at each cell
92 lines of code changed in 4 files:
Compile error fixes
0 lines of code changed in 2 files:
M CCA/Components/Arches/CoalModels/BadHawkDevol.h
M CCA/Components/Arches/CoalModels/ModelBase.h
M CCA/Components/Arches/CoalModels/ConstantModel.h
M CCA/Components/Arches/CoalModels/Devolatilization.h
M CCA/Components/Arches/CoalModels/HeatTransfer.cc
M CCA/Components/Arches/CoalModels/HeatTransfer.h
M CCA/Components/Arches/LU.h
Minor edits
M CCA/Components/Arches/DQMOM.cc
Removing some timings
M CCA/Components/Arches/CoalModels/DragModel.h
M CCA/Components/Arches/CoalModels/DragModel.cc
Removing conversion functions for velocity (Cartesian <--> spherical) and calling them statically from PartVel class
M CCA/Components/Arches/CoalModels/PartVel.h
Making velocity conversion functions (Cartesian <--> spherical) public and static
M CCA/Components/Arches/CoalModels/SimpleHeatTransfer.cc
M CCA/Components/Arches/CoalModels/SimpleHeatTransfer.h
Using vector length()'s instead of the Cartesian <--> spherical methods
Adding verification procedure
M CCA/Components/Arches/CoalModels/KobayashiSarofimDevol.cc
M CCA/Components/Arches/CoalModels/KobayashiSarofimDevol.h
Adding verification procedure
Adding two sets of constants, one from Kobayashi and one from Ubhayakar
Fixing Y1_ value
M CCA/Components/Arches/CoalModels/CoalModelFactory.cc
M CCA/Components/Arches/CoalModels/CoalModelFactory.h
Adding problemSetup method (to be used in future coupled physics calculations)
Adding method to set ArchesLabel (to be used in future coupled physics calculations)
Building models when they are first added to the build map, as opposed to when they are requested for the first time
M CCA/Components/Arches/Arches.cc
Adding call to set ArchesLabel in CoalModelFactory (to be used in future coupled physics calculations)
M CCA/Components/Arches/TransportEqns/ScalarEqn.h
M CCA/Components/Arches/TransportEqns/ScalarEqn.cc
Adding directives header file
M CCA/Components/Arches/Directives.h
Adding new verification directives
A StandAlone/inputs/ARCHES/Verification/verify_devol_heatxfer_models.ups
Adding input files for running new model verification procedures (it also has working moments for 3 internal coordinates and 4 environments)
M StandAlone/inputs/UPS_SPEC/arches_spec.xml
Updating UPS spec
1153 lines of code changed in 33 files:
gcc 4.4 template instantiation fixes for lib uda2vis
0 lines of code changed in 6 files:
autoconf'd (version 2.61) configure to match Justin's checkin of configure.ac
0 lines of code changed in 2 files:
Updates to the histogram code
0 lines of code changed in 2 files:
Updated the out of order scheduler logic.
Updates include:
Post all internally ready tasks recieves, not just the first.
After executing each task, attempt to recieve outstanding MPI without blocking.
Only block when no other work can be perfored.
93 lines of code changed in 1 file:
Need ghost cells for areaFraction
4 lines of code changed in 2 files:
clean up the mess
- added "material" structured array
- added reference comments to the various functions.
- c -> speedSound
- encapsulated the calculation of the problem specific body force into a function
Added code to perform verification using Steffen's MMS equations.
- The main pieces have been written.
- The initial results don't look good.
164 lines of code changed in 1 file:
Updated the fortran library tests so Kraken will configure again.
0 lines of code changed in 2 files:
Forgot to delete the d_newBC object.
0 lines of code changed in 2 files:
Whitespace changes.
Added a debug stream to output a histogram of the schedule queue length.
78 lines of code changed in 1 file:
Small changes to a debug stream.
A few fixes to the random load balancer (it only works for a single level in its current form).
When querying patch assignments operate on the real patch and not the virtual patch. This seems to fix a
crash Jim was seeing when dealing with periodic boundaries.
0 lines of code changed in 2 files:
white space
0 lines of code changed in 2 files:
Fixed a missing ghost cell in previous commit.
2 lines of code changed in 1 file:
Added an area fraction variable that will be used to deal with intrusion and mpmWall boundary conditions in the interior domain. Also added a function to compute the area fraction based on the cell type.
Answers shouldn't change.
239 lines of code changed in 13 files:
Moving the "get" calls so that they only happen once, and not at every cell. This significantly reduces the time spent on these calls.
77 lines of code changed in 2 files:
Code cleanup.
13 lines of code changed in 1 file:
Clean up code and add diagnostic variables
0 lines of code changed in 2 files:
Configure now defaults to the machine's default number of bits.
24 lines of code changed in 2 files:
- overhauled the particle creation section
- now plotting dF and gradient of velG.
findNodesAndWeights()
- looping over the shape function nodes
- added bulletproofing to catch if sum(Ss) != 1.0
Currently setup for the quasi-static compaction problem
with variable resolution, and linear shape functions, PPC=1.
Results look good for this problem
97 lines of code changed in 1 file:
remove warning when not gcc 4.4
0 lines of code changed in 2 files:
One more fix for gcc 4.4 hack
0 lines of code changed in 2 files:
Use template instantiation compiler optimization hack only on gcc 4.4+
19 lines of code changed in 4 files:
Committing configure to match recently committed configure.ac
314 lines of code changed in 1 file:
Need stdio.h for printf... (Caught by gcc 4.4... strange that it wasn't caught by other compilers...)
1 lines of code changed in 1 file:
Fix GCC 4.4 issues with optimizing away needed template instantiation fuctions
37 lines of code changed in 3 files:
Dav: Fixed fortran compiler check. (Wrong autoconf version so not checking in configure right now.
20 lines of code changed in 1 file:
Resolve compile warnings
17 lines of code changed in 1 file:
Model for Deflagration to Detonation Transition based on SimpleBurn and JWL++. It is a work in progress.
986 lines of code changed in 6 files:
Add a needed lib to sub compilation for static build. Fixed comments in configVars to be more informative about libs for static builds.
3 lines of code changed in 3 files:
Change the names of the builders.
Add inferno build slave which does a 32 bit build on 38 cores.
46 lines of code changed in 1 file:
Moving compiler directives for verification into a new header file, Directives.h
33 lines of code changed in 8 files:
Adding DenseMatrix (which utilizes Lapack) to DQMOM class
1087 lines of code changed in 8 files:
added code to evolve lp. This is only partially implemented.
fineNodesAndWeights_gimp2()
- added debugging code to check the answers.
simplified method for computing the zone of influence Lx.
quasi-static compaction problem runs but has problems at the coarse/fine cells
77 lines of code changed in 1 file:
Migrated a few more function calls to the new patch interface.
Deleted the old patch interface.
Removed __New() from the new patch interface calls.
1895 lines of code changed in 376 files:
Renamed the bugged SFC?FORT{Low,High}Index queries to SFC?FORT{Low,High}__Old.
212 lines of code changed in 8 files:
Changes for periodic and no extra cells cases
71 lines of code changed in 4 files:
Readded a constructor that takes the low and high ranges for both the interior and exterior grids.
76 lines of code changed in 1 file:
Changed the specification of the GridSurfaceIterator.
It is now specified by supplying a low and high point for the
grid and an IntVector offset which specifies the number of cells on
the outside or inside (if the offset is negative) of the grid to
visit.
For example, to iterate over the outermost interior layer of a patch you would
use the following:
GridSurfaceIterator(patch_low,patch_high,IntVector(-1,-1,-1))
To iterate over the ghost cells of a patch you would use
GridSurfaceIterator(patch_low,patch_high,IntVector(ngc,ngc,ngc))
20 lines of code changed in 1 file:
Added a new iterator class called GridSurface iterator.
This iterator will iterate over the cells on the surface of a grid which is specified
by the low and high points of an interior and exterior grid. Each cell on the surface
will be touched exactly once.
336 lines of code changed in 4 files:
Reduce the output created by the wait time debug stream.
12 lines of code changed in 1 file:
Shell script for using pump mode for distcc.
7 lines of code changed in 1 file:
Cosmetics
added findNodesAndWeightGradients_gimp2() -for variable resolution grids
This commit give the same results as the stock shape/gradient functions for the
constant resolution quasi-compaction problem.
119 lines of code changed in 1 file:
Matlab code that generates a plot of effective shape and gradient of
the effective shape function when using gimp. Both constant and variable
resolution methods are being plotted along with the numerical differentation
of the shape functions.
307 lines of code changed in 1 file:
Fixing compiler warning
1 lines of code changed in 1 file:
findNodesAndWeights_gimp2() (Multi-resolution version)
- added bulletproofing sum(Ss) = 1.
- added error checking against findNodesAndWeights_gimp().
Quasi-static compaction problem working with this function on a single resolution grid.
60 lines of code changed in 1 file:
Fixed compile errors
5 lines of code changed in 1 file:
Transferring contrl of the model evaluations to the CoalModelFactory.
This will not change answers (no, really, this time I made sure - I fixed the problem with gas-phase source terms).
This transfer will make verification of the model terms easier.
It will also pave the way for implementation of an iterative, coupled particle calculation.
642 lines of code changed in 26 files:
- Set boundary conditions on the accelation.
- fixed bug when settig velocity BC's
Compaction problem results looking good with gimp.
14 lines of code changed in 1 file:
Fixed typo (-) -> (+) when computing last region's max value, gimp only.
Now setting the BC on the first 2 and last 2 nodes in the domain.
moved computeStressFromVelocity() to just below main()
-added bulletproofing to positionToNode()
A particle can't be outside of the computational domain.
findNodesAndWeights_gimp()
-added () around expressions in the conditional statements.
-added bulletproofing.
The shape function weights must sum to 1.0
findNodesAndWeightsGradients_gimp()
-added () around expressions in the conditional statements.
-added bulletproofing.
The gradiend of the shape function weights must sum to 0.0
Compaction problem: issues with the first 2 particles, otherwise
stress distribution looks reasonable
101 lines of code changed in 1 file:
Added a debug stream for outputting waittimes.
8 lines of code changed in 1 file:
Fixing compiler warning
1 lines of code changed in 1 file:
Now interpolating to 3 nodes when using gimp and 2 when using linear shap functions.
- increases the NN in the first/last region
- expand the first & last regions domain.
- added positionToClosestNodes() function, only called by gimp functions
Eliminated colliding Bars problem
numRegions -> nRegions
It doesn't crash, but otherwise it's untested.
127 lines of code changed in 1 file:
Adding the second of two pieces to DQMOM verification procedure.
This check-in adds the ability to compare matrix solutions (&c.) to those calculated in Matlab, for (ill-conditioned) matrices with given condition numbers.
52 lines of code changed in 3 files:
Another allgather hack.
39 lines of code changed in 1 file:
Added another hack to eliminate an allgatherv
38 lines of code changed in 1 file:
Fixing a typo
1 lines of code changed in 1 file:
Adding verification procedures to DQMOM class.
Verification procedures do the following:
1. Verify that matrices A and B are constructed correctly
2. Verify that matrix system AX=B is being solved correctly
The .dat files being added are only for verification of the A and B construction procedure.
Additional .dat files (and .ups files) for verification of AX=B solution process to come later.
896 lines of code changed in 3 files:
added findNodesAndWeightGradients_gimp()
added findNodesAndWeights_gimp()
from MPM documentation, equations 7.16 & 7.17.
I'm able to reproduce plots, 7.3 and 7.4, but not the quasi static compaction problem.
106 lines of code changed in 1 file:
Change the email from line to csafe-homebrew.
Add the -nice option for lowering the priority for regression tester.
2 lines of code changed in 1 file:
Add -nice option to lower the priority of the running jobs.
0 lines of code changed in 2 files:
Tweaked logic to match old logic (my math didn't account for the equals in <=).
7 lines of code changed in 1 file:
Avoid some while loops used in the Data Archiver.
10 lines of code changed in 1 file:
Initialize datatypes correctly.
2 lines of code changed in 1 file:
Multiply body force by massP(ip)
Added back the old method for computing the shape functions to compare against.
(commented out)
Minor tweaks
26 lines of code changed in 1 file:
Fixing but in DQMOMEqn initialization function (and adding brevity to ExplicitSolver statement)
10 lines of code changed in 2 files:
Fixed a small bug in allgather hack.
Release memory sooner.
6 lines of code changed in 2 files:
Added a hack to change Allgatherv to Allgather for testing on Kraken.
Currently Allgatherv on Kraken has some sever performance problems that NICS, Cray, and the MPICH developers are looking into.
80 lines of code changed in 2 files:
Missed an osx fix.
0 lines of code changed in 2 files:
autoconf'd (version 2.61) configure.ac for OSX fixes.
35 lines of code changed in 1 file:
OSX Snow Leopard fixs from Dav. (First OSX to truely support 64 bits.)
M include/sci_defs/osx_testdefs.h.in
OSX Hack isn't used any more, so I removed it. Added the snow leopard #define.
M CCA/Components/Schedulers/OnDemandDataWarehouse.h
No one uses the boundary layer parameter so I removed it.
M CCA/Components/Schedulers/TaskGraph.h
- Removed tabs messing up indentation.
- Made a few params 'const'.
M CCA/Components/Schedulers/TaskGraph.cc
- Mostly cosmetics:
- Added some white space.
- Added some {}.
- Made a few params 'const'.
M CCA/Components/Schedulers/OnDemandDataWarehouse.cc
- An undefined boundary layer was being passed in even though no one uses it, so
default it to 0,0,0.
M CCA/Components/Schedulers/SchedulerCommon.cc
- Use 'const'.
M CCA/Ports/DataWarehouse.h
- Indent.
- In allocateTemporary(), you can't give a reference parameter (boundaryLayer) a default value. Not
sure why this compiles on other machines.
M Core/Grid/Patch.h
- White space cosmetics.
M Core/Grid/Variables/ParticleVariableBase.h
M Core/Grid/Variables/ParticleSubset.h
- Can't have a function differentiated only by return type (and particularly just by 'const' on the return type).
Not sure why other compiles allow this, and/or which function they actually 'link in', but the OSX 10 compiler
doesn't like it.
M Core/Grid/Variables/ReductionVariable_special.cc
- Don't compile some of these things under snow leopard.
M Core/Grid/Variables/ParticleVariableBase.cc
- Cosmetics.
M Core/Thread/Thread.cc
- Fix a warning for John. ;)
M Core/Disclosure/TypeUtils.cc
- 64 bit functions that become multiply defined if left in for Snow Leopard.
M Core/Util/Endian.cc
- White space!
M StandAlone/sub.mk
- Need Core/OS to link.
M StandAlone/tools/puda/varsummary.cc
- 64 bit overloaded function fix for OSX snow leopard.
M configure.ac
- OSX snow leopard fixes.
- NOTE: Not checking in configure because I don't have the correct autoconf version on this
machine. I will fix this in the next commit (coming shortly).
127 lines of code changed in 28 files:
-added quasi-static compaction problem.
-wrapped previously committed debugging code in problem_type conditionals
-now plotting most particle and grid data in plotResults
findNodesAndWeights()
removed old method for computing the shape functions
findNodesAndWeights_old()
nuked
* quasi-static compaction working on a single level with same parameters as
described in GIMP paper.
185 lines of code changed in 1 file:
Dav: Indention is good too. :)
16 lines of code changed in 1 file:
Removed the extra initialization and moved the boundary condition evaluation until after the model is computed.
Also put back a default boolean for particle velocity model.
10 lines of code changed in 1 file:
Reverting CoalModels commit
860 lines of code changed in 19 files:
Cosmetics. Whitespace is your friend. At least my friend. :)
2 lines of code changed in 1 file:
Fix checkpoint UPS verification logic, update comments to be clear that we are dealing with 'seconds'.
39 lines of code changed in 3 files:
Getting rid of smoothTfield (was causing problems)
Adding SimpleHeatTransfer to arches UPS spec
0 lines of code changed in 2 files:
Reverting some previous changes - forgot that we have to initialize abscissas to constant (0.0) in some cases
0 lines of code changed in 1 file:
Changing the default of particle velcity BCs to be the same as gas velocity BCs
(the inlet velocity should be the same as the gas by default)
(this needs to be fixed so that the inlet boundary condition velocity will be the same as the gas, but the outlet/wall/etc boundary condition velocity will be 0.0)
1 lines of code changed in 1 file:
Adding initialization of smooth temperature field back in (removed accidentally)
6 lines of code changed in 2 files:
added mass_L to the variables that can be spewed out.
1 lines of code changed in 1 file:
Removing comments unrelated to BadHawkDevol class (oops)
0 lines of code changed in 1 file:
1. Changing structure of DQMOM coal model evaluations - this will not change the answers!
Instead of ExplicitSolver calling sched_computeModel for each model, it calls the CoalModelFactory once
The CoalModelFactory then calls sched_computeModel (and computeModel) for each model
This will allow for the later implementation of an iterative method for model calculations
2. Adding Devolatilization and HeatTransfer base classes
The "HeatTransfer" class is no longer a specific model, it s a virtual/base class for any kind of heat transfer model
The existing HeatTransfer class has been moved to SimpleHeatTransfer
This is in anticipation of future heat transfer models being created
Additionally, the HeatTransfer and Devolatilization virtual/base classes implement methods that will be useful when implementing an iterative method for coal particle calculations
(i.e. the iterative multiphysics coal model used in Glacier)
3. Adding Badzioch and Hawksley model back in (but not compiling it, since it's still an empty class)
4. Adjusting the check for model types in Arches.cc
1951 lines of code changed in 20 files:
Adding bullet-proofing to make sure we aren't setting all DQMOM abscissas to the same value
10 lines of code changed in 1 file:
Rearranging construction of A and B matrices (to make verification implementation easier)
155 lines of code changed in 2 files: