committing updated version of Arenisca including particle removing algorithm with fixed bug related to numerical cut-off
170 lines of code changed in 2 files:
Removed dependence on CUDA SDK (cutil headers and libs).
CUDA support is now enabled with a single option:
--with-cuda=DIR (CUDA installation is in DIR)
Also generated new configure script.
100 lines of code changed in 8 files:
Error handling wrappers for CUDA runtime and driver API calls.
Working on independence from CUTIL and other SDK related headers/libs.
173 lines of code changed in 2 files:
Criminy. Adding a cerr to let any users of findCellAndNodes27 know that it is
incorrect.
3 lines of code changed in 1 file:
Backed out Jim's changes. He's sick and this isn't helping.
39 lines of code changed in 2 files:
sched_Coarsen_Q()
- added requires statement for data on the fine level.
1 lines of code changed in 1 file:
Removing findCellNodes27(...) as it is incorrect and, fortunately, unused.
0 lines of code changed in 2 files:
added BurnRate regression test, which starts from a restart.
8 lines of code changed in 1 file:
Fixing a mistake in an earlier commit. This should fix the failing rt
test.
3 lines of code changed in 1 file:
Modified JWL EOS in ViscoScram so pressure can no longer go negative. Input file changed for regression test -Jacqueline
183 lines of code changed in 2 files:
Fix compiler error introduced with r48817 commit.
2 lines of code changed in 1 file:
Adapt ARCHES and MPM PETSc interfaces to use PETSC v3.2.
This is a step toward utilizing PETSc GPU interface in PETSc v3.2.
169 lines of code changed in 3 files:
This script searches the output file every N minutes for
an increase in the timestep number. If the timestep
is not increasing the jobID is killed with a qdel command
Manditory inputs:
--jid <job id that will be killed>
--out <name of the outputfile to monitor>
--interval <interval in minutes between searches>
* needs to be tested.
80 lines of code changed in 1 file:
For the soot model: c3 -> c1 to be consistent with literature.
Added an input file using PETSC for the radiation solver.
4 lines of code changed in 2 files:
fix a typo in the RT for Wasatch that was causing some tests to still do restarts.
15 lines of code changed in 1 file:
cout -> proc0cout
1 lines of code changed in 1 file:
Allow speciation in the cold flow mixing model.
125 lines of code changed in 2 files:
grid:
== operator more rigorous test. Patch cells and physical domain
must be the same.
RMCRT
fixed grid comparison conditional.
Thanks again Dav.
19 lines of code changed in 2 files:
problemSetup()
turned on "are the grids the same" bulletproofing.
Thanks Dav.
3 lines of code changed in 1 file:
Kraken compiler warnings
2 lines of code changed in 1 file:
Quieted compiler warnings with virtual destructors.
12 lines of code changed in 2 files:
temporarily turn off restart on the wasatch tests that are failing to conserve memory. The issue seems to go deeper into Uintah than I had though.
12 lines of code changed in 1 file:
fixed bigmac compile.
roundup tabs
formatting -Dav
31 lines of code changed in 1 file:
Keeping consistent with the old soot model, don't allow the model to
work until T > 1000K.
5 lines of code changed in 1 file:
The user can now specify that abskg, temperature and cellType
be initialized from a previously run uda. For example
RMCRT_Test can start off using data from an Arches uda.
Add:
<!-- initialize abskg and temperature using a previously run uda -->
<initalizeUsingUda>
<uda_name> methaneplumeexplicit.uda.000 </uda_name>
<timestep> 25 </timestep>
<abskg_varName> abskgIN </abskg_varName>
<temperature_varName> tempIN </temperature_varName>
</initalizeUsingUda>
to the input file. See
inputs/Examples/RMCRT_udaInit.ups
for an example.
187 lines of code changed in 2 files:
A litle formatting change on the soot model.
Update the 4x4 methane case so that it uses the new soot model. Also
fixed several errors in the input file.
3 lines of code changed in 1 file:
Adding an empirical soot model that computes absorption coef for
radiation and soot volume fraction.
426 lines of code changed in 6 files:
Prep work for initializing abskg & temp from previously run Arches udas.
Ray:
initProperties()
- this task modifies abskg & temperature. It no longer computes sigmaT4
- deleted time_sub_step related code. This was implemented for Arches
and Arches isn't using this task.
sched_CoarsenAll()
- passing in modifies (abskg & sigmaT4) bools.
Arches/RMCRT.cc
- updated sched_coarsenAll() parameter list
RMCRT_Test.cc
scheduleTimeAdvance()
- CarryForward cellType, color and abskg at top of timestep.
This simpilfies everything.
DataOnion
- scheduling sigmaT4 & setBoundaryConditions tasks.
CoarseLlevel approach
- scheduling sigmaT4 & setBoundaryConditions tasks
This mimics the task scheduling in Arches:RMCRT.
pseudoCFD()
- highlighted a bug that needs to be addressed.
The RMCRT_ML test will fail. The answers are the same but
inconsistent material sets for variable abskg in checkpoints.
88 lines of code changed in 4 files:
add rk3 momentum test, multipatch, mpi. use modUPS to modify a base input file. NOTE: you cannot use hyphens in the modified ups filename. also, don't use quotation marks on the modified ups filename in the test definition. here's the documentation for that...
14 lines of code changed in 1 file:
add restart to RT for momentum transport with bcs. this required modification of input file for said test. it will break the RT. this is expected.
2 lines of code changed in 1 file:
add restart to a few bc-scalar tests. this required modification of the input file and hence will break the RT. this is expected behavior. also allow for specification of additional bcs for transport equations.
40 lines of code changed in 3 files:
add restart to some RT periodic test problems.
14 lines of code changed in 1 file:
update some documentation on BCDataArray to make it a bit clearer.
16 lines of code changed in 2 files:
fix varlabel creation for Wasatch restarts when dealing with extracells in Wasatch. this now works with both extracells and ghostcells. NOTE: we're assuming a single ghost layer at the moment.
5 lines of code changed in 1 file:
Populating boundary fluxes using map,
rather than Stencil7. Once verification
testing is complete, will need an additional
regression test.
130 lines of code changed in 2 files:
changed name of input file and made symbolic links to input file in inputs directory -Jacquie
333 lines of code changed in 16 files:
Moving input files from OrderAccuracy to inputs
0 lines of code changed in 2 files:
Some cleanup and fixed compiler warning about unused debug var Nx.
1 lines of code changed in 1 file:
Use "if (!tsk->usesGPU())" instead of "#ifndef HAVE CUDA" to restrict modifies/computes in GPU raytracer.
This is so ThreadedScheduler can be correctly used if Uintah is compiled with CUDA.
11 lines of code changed in 1 file:
Restrict what is modified/computed in GPU ray Tracer.
Essentially, d_boundFluxLabel was casing an exception to be thrown as CCVariable<Stencil7> is not yet supported in the GPU scheduler. This will be addressed soon.
7 lines of code changed in 1 file:
Added support for "patch_based" regions of interest for the data Onion.
rayTrace_dataOnion()
- removed redundant variable 'face' now using dir.
179 lines of code changed in 2 files:
The user can now use multiple DataAnalysis modules. Previously,
they could only use one.
*this shouldn't change the results.
241 lines of code changed in 13 files:
Added utility task carryForward(). Component can use this to move
a variable from old_dw -> new_dw.
The component must initialize cellType on all levels.
RMCRT no longer computes or modifies celltype.
rayTrace()
- cleaned up confusion with requires/gets
- fixed transpose error in parameters
71 lines of code changed in 3 files:
Fix GPU RayTrace compiler error.
5 lines of code changed in 3 files:
Fixing a few print statements.
3 lines of code changed in 1 file:
Getting celltype for the RayTrace task.
Todd: this gives the same answers as your most
recent commit.
32 lines of code changed in 4 files:
updatesumI()
- pass sigmaT4pi & abskg by reference.
18 lines of code changed in 2 files:
Zero order rate model. Included as a stand-alone model and built in to DDT1.
674 lines of code changed in 6 files:
ups files:
removed <Slice> tag since it's not being used.
Ray.cc:
cleaned up the dataOnion specification. The user can specify a ROI_extents as either
being 'dynamic', 'fixed' or 'patch_based.' patch_based has not been
implemented yet.
57 lines of code changed in 2 files:
Added the ability to specify <CommonTags> outside of the main ups_spec.xml file.
You do this by just creating another <CommonTags></CommonTags> block and placing
more common tags in there. Note, however, that all 'Common' tags must have unique
names across the entire UPS spec.
Moved the Wasatch common tags into the qasatch_spec.xml file. (Also indented a few
of the other tags in this file and removed tabs.)
52 lines of code changed in 1 file:
Added multilevel and reflection RMCRT tests.
5 lines of code changed in 1 file:
Ray.cc
setBoundaryConditon()
- if there are multiple levels back out the temperature from sigmaT4
set the boundary conditions, then compute sigmaT4 in extraCells.
On a single level just use the component's temperature field.
Changed user inputs for dataOnion and RMCRT_coarselevel algorithms
RMCRT_Test:
Removed AMR mesh generation from initialization. Nice idea but never used.
Changed user inputs for dataOnion and RMCRT_coarselevel algorithms
now using an int & enums instead of two bool flags.
ups files:
Now specify the RMCRT algorithm used via
<algorithm type='dataOnion'>
<abskg_threshold> 0.1 </abskg_threshold>
<halo> [4,4,4] </halo>
</algorithm>
or
<algorithm type='RMCRT_coarseLevel'>
<orderOfInterpolation> 1 </orderOfInterpolation>
</algorithm>
101 lines of code changed in 4 files:
consolidated gpu & cpu constructor code into a common function.
spaces
applied Roundup to tabs.
59 lines of code changed in 2 files:
remove a few assertions.
3 lines of code changed in 1 file:
Remove convective fluxes from scalability test.
This will break the regression tests on Wastach.
20 lines of code changed in 4 files:
Updates (corrections) to the combustion efficiency calculator.
49 lines of code changed in 1 file:
The user now has the ability to specify the orientation of the tiff image.
63 lines of code changed in 1 file:
Fix uninitialized d_boundFluxLabel in GPU ray tracer.
Need to devise a single CTOR solution, as only C++11 allows a constructor to delegate work to another constructor in the same class.
2 lines of code changed in 2 files:
initialize task phases.
7 lines of code changed in 3 files:
silence more compiler warnings
3 lines of code changed in 3 files:
Silence a compiler warning.
2 lines of code changed in 1 file:
remove some superfluous inclusions.
0 lines of code changed in 2 files:
Allow rectangle-bc-objects to be specified on any minus/plus face, not only those that lowest or highest. Also, fix the way we identify if a boundary point is inside a rectangle. The previous method consisted of adding an arbitrary thickness to the rectangle which failed to work in Wasatch when NOT using extraCells.
41 lines of code changed in 2 files:
Now passing celltype label to Ray.cc to inform it of intrusion boundary
conditions.
Updated example component so that celltype is initialized through
geom_object specification from the input file.
Added a new input file (RMCRT_test_1L_wIntrusion.ups) as an example of
specifying an intrusion geometry.
Todd: Please let us know if there is anything you don't like here. We
are happy to iterate.
Jeremy & Isaac
76 lines of code changed in 6 files:
Fixed a few bugs in the efficiency calculator.
Updated the T-channel to use the new boundary conditions.
5 lines of code changed in 1 file:
* added reference density location to periodic driven cavity problem
* Intrusions now store their user-defined name
* Added almgren initialization function for velocity.
* Changed the way mixture fraction and density are handled on inlets. Unlike Arches legacy code, the mixture fraction and the density are now consistent when using the new boundary conditions.
This change will cause these tests to fail:
:xplus_scalar_test: test failed comparison tests
:yplus_scalar_test: test failed comparison tests
:zplus_scalar_test: test failed comparison tests
:xminus_scalar_test: test failed comparison tests
:yminus_scalar_test: test failed comparison tests
:zminus_scalar_test: test failed comparison tests
149 lines of code changed in 7 files:
rename the physicscoefficient to PrecipitationBasicExpression. keep it as a basic expression but separate it from other general purpose basicexpressions - hence the nomenclature. Alex, please update your input files for the warches problems.
2 lines of code changed in 1 file:
Add GPUStats dbg stream to GPUThreadedMPI scheduler.
export SCI_DEBUG=GPUStats:+ will give GPU allocation and copy stats per device, per process.
25 lines of code changed in 1 file:
-compiler warning
Formatting:
- 2 space indentation
- vertical alignment
- align terms in equations.
- no tabs.
- comments
This should not change the answers
271 lines of code changed in 1 file:
More descriptive file name.
4 timesteps.
Added 1-Level, RMCRT benchmark 1 test, to Arches nightly & local RT.
3 lines of code changed in 1 file:
Sigh. Fix TWO compiler warnings that my fix of the previous compiler warning
kicked up.
2 lines of code changed in 1 file:
Fix compiler warning.
1 lines of code changed in 1 file:
Fix crashing dbg tests by adding a requires for g.mass. Also, only schedule
cohesive zone tasks if they are actually used.
5 lines of code changed in 1 file:
cd to order_of_accuracy before executing the next component's tests.
2 lines of code changed in 1 file:
Documented external library dependencies of particle2tiff.
4 lines of code changed in 1 file:
yanked tabs.
7 lines of code changed in 1 file:
rmcrt_test.ups:
- added temperature boundary conditions for RMCRT
SourceTerms/RMCRT.cc
- global change color -> temp
- no longer initializing abskg or temperature, that's handled upstream
- added extraSetup() step in problemSetup() task to set the temperature label
- updated argument linst in sched_sigmaT4
- now setting boundary conditions on abskg & sigmaT4 after
compute_sigmaT4() task.
- no longer setting boundary conditions in radProperties()
ray.cc/h
sigmaT4()
-Pass in includeEC, if you want to include extra cells in the computation.
setBoundaryConditions()
- use a copy of the temperature and setBC on that copy before computing
sigmaT4 in the extraCells.
RMCRT_Test.cc
- moved sheduling of ROI_Extents
- update argument lists
114 lines of code changed in 5 files:
Fix the requirement for a "res" for axisymmetric cylinder ends. Also, don't
issue Todd's warning about geometry exceeding domain in the case of axisymmetric
geometry.
13 lines of code changed in 1 file:
added comments,
formatting
delete scinew'd variables.
32 lines of code changed in 1 file:
Allow control of thread counts for ExprLib and SpatialOps from the input file.
18 lines of code changed in 1 file:
- added user defined min/max clamps for the averaged data.
- user can now average 1 matl or all matls via the -m option
-m [int]
or
-m 'a' or 'all'
145 lines of code changed in 1 file:
fix udaReader problem with ghost zones
127 lines of code changed in 2 files:
- added 32 bit support for images. The computed averages can now be saved
as the pixel values. There's no need to normalize the data.
- cleaned out testing & thrashing code.
- Added support for computing the average over all matls. Thanks Dav!
- Added Matrix3 version of compute_ave()
- reformatting & cleanup
188 lines of code changed in 1 file:
Removed unused code.
0 lines of code changed in 2 files:
A little cleanup.
0 lines of code changed in 6 files:
Incorporation of Andy Tonge changes to allow for input points to be defined on AMR grids.
56 lines of code changed in 2 files:
Updates/improvements to the cohesive zone code. --Ravi
37 lines of code changed in 1 file:
Modifications to conform to new SpatialFieldStore interface.
95 lines of code changed in 12 files:
Add some minor documentation.
19 lines of code changed in 2 files:
Added a new post processing utility that takes an uda
averages the particle data in each cell to the cell-center and
dumps out a stack of tiff files. Still thrashing around.....
Usage: tools/extractors/particle2tiff [options] -uda <archive file>
Valid options are:
-h, --help
-v, --variable: <variable name>
-m, --material: <material number> [defaults to 0]
-tlow, --timesteplow: [int] (sets start output timestep to int) [defaults to 0]
-thigh, --timestephigh: [int] (sets end output timestep to int) [defaults to last timestep]
-timestep, --timestep: [int] (only outputs from timestep int) [defaults to 0]
-istart, --indexs: <i> <j> <k> [ints] starting point cell index [defaults to 0 0 0]
-iend, --indexe: <i> <j> <k> [ints] end-point cell index [defaults to 0 0 0]
-startPt <x> <y> <z> [doubles] starting point of line in physical coordinates
-endPt <x> <y> <z> [doubles] end-point of line in physical coordinates
-l, --level: [int] (level index to query range from) [defaults to 0]
-d, --dir: directory name where all output is kept [none]
--cellIndexFile: <filename> (file that contains a list of cell indices)
[int 100, 43, 0]
[int 101, 43, 0]
[int 102, 44, 0]
----------------------------------------------------------------------------------------
802 lines of code changed in 4 files:
The old ExprAlgebra is removed from function.h and Then new one is modified so that in can be used from input files. For now it just supports operation between same field types to the same field type as it was before when it is used from input file, but internally you can use it without that limit. The support for operations between different field types to different field type in input files will be added in the future commits.
Multiplier expression is replaced with the new ExprAlgebra expression as well.
ScalarTransportEquation.h initialization method is updated.
MomentumTransportEquation.h initialization method is modified so we can initialize velocity or momentum in the input file from now on without ant problem.
53 lines of code changed in 9 files:
Reduce the number of pressure solve iterations in the 2nd and 3rd RK3 integrator. The key to achieving this is to NOT initialize the pressure to 0 in the evaluate method for these stages, because we are using the pressure in the newDW as initial guess for the 2nd and 3rd RK stage. \n
19 lines of code changed in 3 files:
remove obsolete commented code in pressure expression.
0 lines of code changed in 1 file:
Fix bug in RK3 multipatch-mpi pressure solve.
8 lines of code changed in 1 file:
I added ExprAlgebra.h for algebraic operation between two expression in differant fields just for internal usages in wasatch
226 lines of code changed in 2 files:
Rip out a few more ".first" things that make my compiler whine at me.
Jeremy - I like the OperatorSplitChem...
3 lines of code changed in 3 files:
Flux variables now being initialized properly. We may not keep the
Stencil7 format for the flux, but if we do, we'll need this commit.
24 lines of code changed in 1 file:
Handle 'have_tiff' correctly.
3 lines of code changed in 2 files:
add pressure outlet tests to regression tester.
12 lines of code changed in 1 file:
Remove a few dead code fragments (.first)
8 lines of code changed in 1 file:
Added Stencil7 variable label to boundary fluxes.
35 lines of code changed in 2 files:
clean up debug comments, fix compiler warnings
11 lines of code changed in 2 files:
hack in a task to apply pressure boundary conditions after the pressure solve. As of this commit, the pressure-outlet (e.g. specified pressure) is fully functional with non-zero pressure values at the outlet. add 6 tests that exercise the pressure outlet bc in all directions (soon to be added to the RT). Add a jet test case (with co-flow) so that Amir can start testing his variable density solver << std::endl;
288 lines of code changed in 6 files:
Comprehensive udaReader fixes:
o correct domain calculations for face-centered data, including periodic
o calculate domain nestings for each mesh (instead of assuming all meshes are the same as CC mesh)
o only add procId variable if there is a CC or NC mesh for it to live on
o reduce redundant computations on udaReader
o simplify computation of logical bounds
o added getBounds function to utilize only one method to get bounds for any or all patches of any mesh type (previously bounds were manually computed in a variety of different ways)
o remove extraCells from udaData
o add helper functions to PatchInfo to get/set bounds for different mesh types
o comment out all unnecessary debug spew
346 lines of code changed in 3 files:
Fix a bug on the scalability test that was giving false impressions that the monolithic approach was super fast.
11 lines of code changed in 1 file: