fixed memory leak.
4 lines of code changed in 1 file:
computeStressTensor
- changed order of operations in tensorS calculation
computeStressTensorImplicit
- minor rearrangment in the equations for clarification, changing the order
of operations.
- implemented calls do deviatoricStressModel()
## This will change the answers in input files that use ElasticPlastic.
47 lines of code changed in 1 file:
Use Exceptions, not exit(1).
Tabs are the devil
Comment out some unused variables, reduce warnings.
Should not change answers.
164 lines of code changed in 1 file:
This should fix the failing bc-test-mixed RT for Wasatch.
1 lines of code changed in 1 file:
1. fix a bug in the evaluate that was causing a segfault.
2. implement some assertions.
11 lines of code changed in 1 file:
Correct a sign error in the MonolithicRHS expression
6 lines of code changed in 1 file:
Fix a bug on STATE_N vs STATE_NONE for RHS in scalability test.
5 lines of code changed in 1 file:
Combine convective fluxes into monolithic approach
16 lines of code changed in 1 file:
First cut at monolithic RHS - still needs a lot more work.
389 lines of code changed in 4 files:
Updated Boundary Flux documentation to include
verification results and ray convergence analysis
430 lines of code changed in 6 files:
Fix for (copy pasted) memory leak.
4 lines of code changed in 1 file:
Rip out a few strange code fragments that were generating warnings on my builds.
14 lines of code changed in 2 files:
Fix a bad bug related to interpolating area fractions.
Rework ScalarRHS quite a bit to reduce the number of loops, etc.
74 lines of code changed in 1 file:
Replaced existing deviatoric stress model with a "factory" which includes HypoElastic &
HypoViscoElastic models. The main reason was to support
std::vector<ParticleVariable<Matrix3>
for each particle needed by the HypoViscoElastic model.
This should not change the answers.
932 lines of code changed in 12 files:
Copy the sus output and the compare_uda output to a directory so it can be uploaded
to a webserver.
7 lines of code changed in 1 file:
Remove the pressure outlet tests until we get the process_after_evaluate for pressure properly handled.
0 lines of code changed in 1 file:
runSusTests.py
- redirect output to memory leak log file
mem_leak_checkVarLabels
- restrict is searched for when creating and destroying variables
3 lines of code changed in 3 files:
fixed symbolic link to tools
0 lines of code changed in 2 files:
added environmental variable WHICH_TESTS when creating "do<component>Test" scripts
0 lines of code changed in 2 files:
Input file for postprocessing was wrong - Jacquie
13 lines of code changed in 1 file:
computeStressTensor()
removed a sanity check.
1 lines of code changed in 1 file:
Fix memory leak. Same code as previous commit I made. Don't need a pointer, can just use a var.
4 lines of code changed in 1 file:
updated post processing tools for determining the burn rate -Jacquie
12 lines of code changed in 9 files:
computeFlowStress()
- changed d_CM.K*pAlpha[idx]; -> d_CM.K*state->plasticStrain; -Scott
4 lines of code changed in 1 file:
Fix the failing RT dbg tests. An unused access to the pressure-coef-matrix leaked into the code yesterday and forgot to remove it. NOTE: when accessing Stencil7 or Stencil4 Coefs, cell indices are global - so be careful in using patch offsets as is the case with all Wasatch spatialfields where indexing is local.
1 lines of code changed in 1 file:
turn BC setting for pressure.
8 lines of code changed in 1 file:
Turn off restarts for intrusion test until the memory bug is fixed.
2 lines of code changed in 1 file:
Add regression tests for pressure outlet boundary conditions. also add 3D-lid-driven-cavity.
11 lines of code changed in 1 file:
Fix memory leak...
4 lines of code changed in 1 file:
Add support for pressure-outlet boundary condition.
111 lines of code changed in 4 files:
build
- chgrp and chmod recursively after configure command has been executed
replace_gold_standard
- no longer sending John spam.
2 lines of code changed in 2 files:
replace a few field assignments with nebo-style assignment operators
50 lines of code changed in 4 files:
Fix a bug in type definition.
2 lines of code changed in 1 file:
Missing ".ups" in script.
2 lines of code changed in 1 file:
Added the ability for sources to do extra setup if needed after the
property stuff is determined.
11 lines of code changed in 2 files:
Check that timeintegrator exists before parsing the CARRY_FORWARD expressions. This is needed when running warches.
1 lines of code changed in 1 file:
Remove UINTAHSHARE and uintahshare.h include files.
387 lines of code changed in 253 files:
* Still need some level code for domain extents
* Fix device-side RNG seed
* Save restricted pointer usage for later optimizations
* Cleanup and commenting
79 lines of code changed in 2 files:
Fix a few issues that were causing a compiler to complain.
62 lines of code changed in 2 files:
Half space and explosion with cam clay.
321 lines of code changed in 1 file:
Verification tests.
322 lines of code changed in 3 files:
Explanation of boundary flux method.
940 lines of code changed in 3 files:
Rip out the superfluous "AnyExpression" tag.
18 lines of code changed in 1 file:
Eliminate GPU ray tracer code that is now being handled by the infrastructure (GPU scheduler).
0 lines of code changed in 1 file:
Fix method resolution problems that were occurring on some compilers.
131 lines of code changed in 1 file:
Fix a bug in ProblemSpecReader
1 lines of code changed in 1 file:
another shot at making the compilers happy - not sure why my compiler is letting this slip through...
1 lines of code changed in 1 file:
add a missing include that choked some compilers.
1 lines of code changed in 1 file:
Get rid of the "temperature transport equation" which was not ever really completed anyway.
0 lines of code changed in 4 files:
Consolidate convective flux expression builder to reduce code duplication & eliminate it from the Scalability test.
Also improve some error trapping, and modify the ConvInterpStringMap handling to make it a boost::bimap for easier backward access.
192 lines of code changed in 6 files:
Allow users to specify reference pressure location and value. update all input files in need of this.
71 lines of code changed in 5 files:
computeDeltaGamma()
- changed iteration loop tolerance. --Scott
This will change the following answers:
__________________________________
S U M M A R Y
Failed: IMPM-opt tests
:adiCuPTW01s296K: test failed comparison tests
:adiCuPTW01s296K: (restart) test failed comparison tests
Failed: MPM-opt tests
:adiCuJC4000s696K: test failed comparison tests
:adiCuJC4000s696K: (restart) test failed comparison tests
:adiCuMTS4000s696K: test failed comparison tests
:adiCuMTS4000s696K: (restart) test failed comparison tests
:adiCuPTW4000s696K: test failed comparison tests
:adiCuPTW4000s696K: (restart) test failed comparison tests
:adiCuSCG4000s696K: test failed comparison tests
:adiCuSCG4000s696K: (restart) test failed comparison tests
:adiCuZA4000s696K: test failed comparison tests
:adiCuZA4000s696K: (restart) test failed comparison tests
:test_corrug_plate: test failed comparison tests
:test_corrug_plate: (restart) test failed comparison tests
:test_cyl_pene_no_ero: test failed comparison tests
:test_cyl_pene_no_ero: (restart) test failed comparison tests
:test_gurson_beckerdrucker_mts: test failed comparison tests
:test_gurson_beckerdrucker_mts: (restart) test failed comparison tests
:Charpy: test failed comparison tests
:Charpy: (restart) test failed comparison tests
1 lines of code changed in 1 file:
Move some parsing things from the ScalarTransportEquation to ParseEquation to make them more accessible elsewhere. These were also reworked a bit to eliminate some redundant code, etc.
Simplify a few things when dealing with expressions that map into the ScalarRHS expression.
Minor cleanup.
454 lines of code changed in 13 files:
Fixing a problem with the swirl BC.
5 lines of code changed in 1 file:
Fix a few crazy mistakes that slipped through into my last commit.
7 lines of code changed in 2 files:
Ensure that any CARRY_FORWARD fields are present in the initialization factory.
58 lines of code changed in 4 files:
Minor changes & add a citation for the 3D Taylor vortex problem.
13 lines of code changed in 1 file:
Handling non-convergence better
26 lines of code changed in 1 file:
Restoring Merrick's correlation
8 lines of code changed in 1 file:
Hopefully this will fix the debug nightly builds that use the sci malloc stuff.
11 lines of code changed in 1 file:
Fix failing test - make sure that hypre solver is only needed when we are solving momentum equations. it seems that Uintah always defaults to CG Solver...
1 lines of code changed in 1 file:
revert recent commit as it committed files not ready for repository yet.
13 lines of code changed in 6 files:
Fix failing test - make sure that hypre solver is only needed when we are solving momentum equations. it seems that Uintah always defaults to CG Solver...
72 lines of code changed in 6 files:
Make the exception thrown in my previous commit a runtime_error instead of internal error.
5 lines of code changed in 1 file:
Throw an exception when a solver other than Hypre is used with Wasatch. This is because of the Stencil4 stuff. The other solver interfaces are not Stencil4 friendly.
4 lines of code changed in 1 file:
Added capability to compute radiative fluxes at
boundaries. Passed sanity checks, but still needs
a little more verification before it's ready
for regression testing.
187 lines of code changed in 2 files:
Changes made to Burn Rate scripts for more accurate results -Jacquie
937 lines of code changed in 19 files:
added RMCRT_VR test.
6 lines of code changed in 1 file:
Missing comma.
1 lines of code changed in 1 file:
Turning on a few new tests.
4 lines of code changed in 1 file:
Function calls needed reordering
1 lines of code changed in 1 file:
Adding restart capability for the new Intrusions.
101 lines of code changed in 8 files:
Fix a vector initialization problem.
3 lines of code changed in 1 file:
Add option to add volumetric expansion behavior to Cp calculation.
26 lines of code changed in 3 files:
Eliminate a redundant declaration of some malloc related routines that was causing errors on a new compiler.
5 lines of code changed in 1 file:
Eliminate one of the TaskInterface constructors to reduce code maintainence.
3 lines of code changed in 3 files:
Fix an improper initialization.
57 lines of code changed in 1 file:
Fix a bug that arose in the situation where a RHS node was an internal node in a graph.
12 lines of code changed in 2 files:
fixed a bug in Arenisca.cc related to rotation
7 lines of code changed in 1 file:
remove a dw exist call in scrub
13 lines of code changed in 2 files:
a missing file for last commit of using gcc atomic operations
87 lines of code changed in 1 file:
Use gcc atomic operations for atomic counters when available.
16 lines of code changed in 8 files:
Quiet the RMCRT-GPU tests until failing UDA comparisons can be resolved.
2 lines of code changed in 1 file:
Allow users to specify body force in momentum equations. Add density to the partial momentum RHS - it is needed by the body force source term (rho * body_force). Add hydrostatic pressure test in a confined volume. Fix some formatting and typos here and there.
63 lines of code changed in 7 files:
Use gcc style atomic operations for reference counting when available.
486 lines of code changed in 5 files:
Indentation continues to be my friend (uintahshare.h) - Note, do we still need this? I think that 'share' stuff is for Windows and I think we pulled most of the windows build stuff out...? Added Constant*M operator to complement M*Constant operator.
15 lines of code changed in 2 files:
This is a fix for when INF ghost cells are requested such as in RMCRT, e.g. tsk->requires(abskg_dw, d_abskgLabel, gac, SHRT_MAX)
SHRT_MAX was used to represent INF ghost cells, which caused issues when allocating
space on the GPU, e.g. when getWindow()->getData()->size() was called with a patch
size of (8,8,8) and EC=1, then (SHRT_MAX, 10, 10) was returned.
34 lines of code changed in 3 files:
Adding some logic to distinguish betweem massflow inlets and velocity
inlets for the flat-profile intrusion.
9 lines of code changed in 1 file:
Removing this option for now.
0 lines of code changed in 1 file:
Updated the Kayenta wrapper with the addtional input parameters for joints. Now the wrapper is consistent with the kayenta trunk for UI's and ISV's
34 lines of code changed in 1 file:
Reverted to CUDA 4.0 after major issues with CUDA 4.1 compiler. Now cudaHostRegister(h_reqData, nbytes, cudaHostRegisterPortable) doesn't work?? Commenting out for now. Implicit page-locked buffer is still created for H2D async transfers, just not quite as fast as registering preallocated host memory.
18 lines of code changed in 1 file:
* Clean up and simplify GPU ray tracer
* Add -G0 to NVCC_FLAGS when debug is enabled
* update configure script
* Add some other test tasks to gpu-scheduler-test component
200 lines of code changed in 6 files:
minor changes to build of wasatch 3p libs
1 lines of code changed in 1 file:
remove some dead code.
6 lines of code changed in 1 file:
Still in testing phase.
821 lines of code changed in 4 files:
Newline at end of file to quite compiler warning.
1 lines of code changed in 1 file:
Accidently removed a bunch of local RT tests in a previous commit.
16 lines of code changed in 1 file:
Cleaning up some warnings. (Thanks, James)
7 lines of code changed in 3 files:
Hooray! Fixed the problem with returning a Vector from
getIteratorBCValueBCKind. Removed the "vecvalue" option from the BCType
spec.
66 lines of code changed in 3 files:
Add reference pressure specification for Wasatch as well as the ability to construct full 3D Laplacian for the pressure in 2D problems. This has important implications on the condition number of the pressure coefficients matrix and therefore the tolerance that Hypre will be able to achieve. Also clean up some obsolete cout statements, add a couple of ifdefs, and remove some switch cases that are obsolete.
140 lines of code changed in 8 files:
Clean up a bunch of sloppy, non-standard code.
For a particularly egregious example, have a look at src/Cores/Basis/Locate.h !
Most of this was due to a new compiler that is catching a lot of things that slipped through the cracks before.
752 lines of code changed in 21 files:
Giving up: Having trouble when the bc_value returned from
getIteratorBCValueBCKind is a vector. Reverting the hack I had
implemented earlier.
23 lines of code changed in 1 file:
added subcycling for F to Arenisca
319 lines of code changed in 1 file:
Incorporate periodicity (2*Pi) into the Taylor-Green vortex MMS (2D & 3D) so that we can specify domain sizes more accurately as integers. Interestingly, this changed the answers on the Taylor-Vortex problems. Solutions are still correct though within discretization error. Will update gold standards tomorrow morning.
6 lines of code changed in 1 file:
If a particle is going to be deleted, set the D tensor to zero so as not to stress the model evaluation.
1 lines of code changed in 1 file:
Implement subcycling for the calculation of F. Also, removed tabs, and mothballed the viscousStress update for localized particles as we're nowhere near ready
for that kind of Hail Mary at this point.
130 lines of code changed in 1 file:
Added Charpy.ups in the inputs/MPM directory and made that as RT for Elastic Plastic Constitutive Model in MPM.py
2 lines of code changed in 1 file:
add some error trapping diagnostics
6 lines of code changed in 1 file:
Added better error checking for the reference pressure point.
Removed "vecvalue" from boundary condition specification. This should
now just work with simply "value".
89 lines of code changed in 3 files:
Simplify some things for the ScalabilityTestTransportEquation & its parsing
49 lines of code changed in 3 files:
Dump version information for TabProps in output.
5 lines of code changed in 1 file:
Make Wasatch use the Linear TabProps build from now on.
Replace table in rtest with a linear interpolant one.
3 lines of code changed in 1 file:
fix some comments
2 lines of code changed in 1 file:
remove some unnecessary code.
1 lines of code changed in 1 file:
Sigh!
added delete defState
ASSERT(trialS_tmp == trialS) -> ASSERT(trialS_tmp != trialS)
3 lines of code changed in 1 file:
Added rotation of elastic strain.
6 lines of code changed in 1 file:
Added rotational correction. The algorithm is still valid only for small strains.
25 lines of code changed in 2 files:
added test_gurson_beckerdrucker_mts test back into the regression tester
computePlasticStateBiswajit()
- removed ASSERT(stst != 0.0); --Scott
which was causing test_gurson_beckerdrucker_mts test to fail.
4 lines of code changed in 2 files:
Two changes:
* Help Todd sleep at night by removing the "this sucks" suckiness of
copying over the stencil to CCVariable<double> so that it can be
passed into the fortran. The fortran was removed in favor of some
sensible C++.
* Fix a problem with the reference pressure option: Pressure wasn't
being set to the user specified value and neighboring stencil values
weren't being adjusted properly.
171 lines of code changed in 4 files:
Solving Todd's confusion. ;)
2 lines of code changed in 1 file:
Added DeformationState Structure
Added DeviatoricStressModel 'factory'
- The user can select
<deviatoric_stress_model type = "hypoElastic"/>
or
<deviatoric_stress_model type = "hypoViscoElastic"/>
The default is hypoElastic.
Most of the plumbing is roughed in just need to hook up fixtures.
* This should not change the answers
233 lines of code changed in 4 files:
replace_gold_standard: Also email Alan whenever there is a change to a GS (for GPU RT).
3 lines of code changed in 1 file:
Add documentation to the Cylinder initialization in Wasatch.
19 lines of code changed in 1 file:
cleanup
43 lines of code changed in 1 file:
Fixed a bug that ghost cells were not mapped correctly between levels when creating neighborhood on commit 47536
1 lines of code changed in 1 file:
Fix configure with gperftools on linux.
2 lines of code changed in 2 files:
Skip initiateTask if no external dependency.
283 lines of code changed in 2 files:
removed failing tests until they can be dealt with.
2 lines of code changed in 2 files:
Changed d_plasticConvergenceAlgo from an int to a string.
I don't like it but it makes dealing with restarts much easier.
6 lines of code changed in 2 files:
Add cylinder initialization function in Wasatch.
165 lines of code changed in 2 files:
Input files for ONR-MURI simulations.
7752 lines of code changed in 27 files:
Updated the python version warning to reflect path change at chpc.
2 lines of code changed in 1 file:
if python version is < 2.5 then print a warning and exit
1 lines of code changed in 1 file:
computePlasticStateBiswajit()
- Pass matrix3 variables in by reference, especially if you want to modify it!!!
6 lines of code changed in 2 files:
Initialize a few additional fields for the pressure solver to avoid random number popping up in the regression tests. This will break the RT for wasatch. I will update tomorrow morning.
15 lines of code changed in 4 files:
Additions/Fixes to build system CUDA support
* Fix issue with missing debug symbols in device code
* Add flags to compile device code specifically for Fermi architecture (will add Kepler flags later)
* Update configure script
24 lines of code changed in 2 files:
Throw an exception if the user selects Gurson yield condition with radial return.
The user is told to add:
<plastic_convergence_algo> biswajit </plastic_convergence_algo>
to the input file
34 lines of code changed in 3 files:
Corrected signs.
36 lines of code changed in 8 files:
Updated. Some signs may stiff be wrong.
69 lines of code changed in 2 files:
Fix gperftools (google code profiling tools) configure/link on Mac OSX.
169 lines of code changed in 5 files:
Remove some unused variables that creeped in from recent changes to BC processing.
0 lines of code changed in 1 file:
Add lid-driven-cavity tests to the RT. NOTE: we are using a single MPI process for the moment because of inconsistent values still showing up at processor boundaries with multiple MPI processes.
8 lines of code changed in 1 file:
Remove some redundant code.
Minor formatting changes.
93 lines of code changed in 3 files:
add a few more memory regression tests for Wasatch.
7 lines of code changed in 1 file:
Initialize a field to avoid NANs or Random values on out-of-boundary faces.
1 lines of code changed in 1 file:
Forgot to commit changes to a sub.mk file during my last commit...
1 lines of code changed in 1 file:
Remove an unnecessary implementation file
Remove a namespace opening in a header file (bad!)
15 lines of code changed in 2 files:
computePlasticStateBiswajit()
- now returning tensorS instead of Stilde.
- moved
Stilde = Stilde*(state->yieldStress/stst);
into CPS_Biswajit()
changed default plastic convergence plastic algorithm from biswajit to radial return.
* This WILL change the answers.
26 lines of code changed in 2 files:
Remove debugging statements.
0 lines of code changed in 1 file:
Add in configure hooks for RadProps. Sub.mk and include files still need to be updated to use these hooks.
106 lines of code changed in 3 files:
Indentation is your (well, at least my) friend.
3 lines of code changed in 1 file:
Moved the VR solve out of the cell iterator that
contains the divQ solver, and into its own iterator.
This allows VR and divQ to be solved independently.
Also added capability for users to specify a line
or block of radiometers rather than a single one.
VR is ready for regression testing.
142 lines of code changed in 2 files:
Additional possibly useful scripts for processing data and frames into movies. Both are rip offs of scripts that Todd wrote previously.
111 lines of code changed in 2 files:
Allowing for mass flow inlets from intrusions.
19 lines of code changed in 2 files:
Fixed a dependency in the UPS file for the SolidReactionModel as well as fixed the problem setup operation to correctly read child nodes of the SolidReactionModel section of input file.
10 lines of code changed in 1 file:
- Change device function parameters to be restricted pointers (alleviate potential aliasing problems).
- Change int to size_t for byte calculations passed to CUDA Runtime API calls (GPUThreadedMPIScheduler).
- Make UDA comparison exact on RMCRT GPU test.
17 lines of code changed in 4 files:
Add RMCRT_1L test to nightly GPU RT (use threaded GPU scheduler). UDAs should compare.
4 lines of code changed in 1 file:
Fixing tabprops for last commit.
2 lines of code changed in 2 files:
remove a cout statement for Wasatch_in_Arches.
0 lines of code changed in 1 file:
Completed initial version of Arches GPU ray tracer. Need to work out minor correctness issue, then focus on optimization.
277 lines of code changed in 3 files:
Variables with arbitrary naming conventions can now be mapped to
specific CFD roles either from the input file:
<VarID>
<var label="my_label" role="temperature"/>
<VarID>
or from individual models using ArchesLabel::setVarlabelToRole.
Varlabels are returned for specific roles from ArchesLabel using
ArchesLabel::getVarlabelByRole. This
will allow models that use the CFD quantities to easily identify needed VarLabels
if the model developer chooses not to make the specification in the
model and its input spec.
Note that permitted CFD roles are defined in ArchesLabel constructor and
should be replicated in the spec.
126 lines of code changed in 11 files:
small logic change in ThreadedMPI2 scheduler
10 lines of code changed in 1 file:
Documents that explain the models. Always work in progress.
11912 lines of code changed in 10 files:
Improved decentralized multi-thread scheduler.
No more busy locking on all threads.
100 lines of code changed in 4 files:
Added option for camclay and now allowing compile of stuff in Biswajit.
5 lines of code changed in 2 files:
Updated to make consistent with other recent changes.
9 lines of code changed in 2 files:
filename changes
SCGPlastic -> SCGFlow
ZAPlastic -> ZAFlow
ZAPolymer -> ZAPolymerFlow
MTSPlastic -> MTSFlow
PTWPlastic -> PTWFlow
JohnsonCookPlastic -> JohnsonCookFlow
IsoHardeningPlastic-> IsoHardeningFlow
PlasticityModelFactory -> FlowStressModelFactory
PlasticityModel -> FlowModel
6055 lines of code changed in 39 files:
Variable & class name change
*.ups: plasticity_model -> flow_model
*.cc/h: d_plastic -> d_flow
PlasticityModel -> FlowModel
PlasticityModelFactory -> FlowModelFactory
IsoHardeningPlastic -> IsoHardeningFlow
JohnsonCookPlastic -> JohnsonCookFlow
MTSPlastic -> MTSFlow
PTWPlastic -> PTWFlow
SCGPlastic -> SCGFlow
ZAPlastic -> ZAFlow
ZAPolymer -> ZAPolymerFlow
This should not alter the answers. Renaming filenames is next.
536 lines of code changed in 20 files:
Fix swirl to work with domains that don't originate at (0,0,0).
Update square_boiler_1.ups to look more like a boiler (ifrf).
8 lines of code changed in 1 file:
sperate partcile varDB from grid&level varDB
108 lines of code changed in 2 files:
Moderately more informative error message.
1 lines of code changed in 1 file:
In some debugging output, use the particleID, not just the loop index, to
identify the particle in question.
3 lines of code changed in 1 file:
Added, in an #ifdef, the option of using a subcycling algorithm to more
accurately compute the deformation gradient.
57 lines of code changed in 1 file:
Enable restart on Wasatch.
81 lines of code changed in 3 files:
check usesMPI() instead of Task::OncePerProc to determin if a new
synchronization phases need to be created to avoid deadlock.
A task need to set usesMPI(true) if itself explicitly uses MPI.
17 lines of code changed in 8 files:
computeTangentModulus()
Removed code and replaced it with an exception.
These functions are never called.
30 lines of code changed in 6 files:
A little cleanup on the coldFlow model.
5 lines of code changed in 3 files:
Reverted changes made on r48480
199 lines of code changed in 2 files:
BurnsChriston Class:
Global: demarcate global vars with "_"
ProblemSetup()
- added bulletproofing
- using getWithDefault instead of get
computeProps()
- fallout from global variable name change
Now runs properly.
24 lines of code changed in 1 file:
added order of accuracy tests for ARCHES::RMCRT_bm1
0 lines of code changed in 2 files:
Using mkpath instead of mkdir. This allows subdirectories
within a component.
3 lines of code changed in 1 file:
added optimization log
baseline timings
10 lines of code changed in 1 file:
Update the treatment of Boundary Conditions for Staggered fields when using Neumann boundary conditions. With this more rigorous formulation, the ghost/extra cell for Neumann conditions is set using a wide finite-difference stencil. In this case, we are able to update (i.e. advance in time) the staggered field at the boundary. This change will break all Wasatch BC regression tests that transport staggered fields and have Neumann BCs applied in the staggered direction (3 tests). This commit also works when using extra cells with Wasatch.
72 lines of code changed in 1 file: