September 2014 Commit Log

Number of Commits:
107
Number of Active Developers:
12
tsaad 2014-09-30 16:18 Rev.: 52524

repair the failing Wasatch test for interior bcs. also, output more timesteps in that test. this will break the RT.

1 lines of code changed in 1 file:

jthornoc 2014-09-30 16:06 Rev.: 52523

Fixing a bug that crept in for handoff files.
Also, allowing tabulated BC's from "FromFile" BCs.

13 lines of code changed in 2 files:

jthornoc 2014-09-30 16:06 Rev.: 52522

Warnings.

3 lines of code changed in 4 files:

u0407585 2014-09-30 15:26 Rev.: 52521

Fixed bug that prevents restart in this devolatilization model.

1 lines of code changed in 1 file:

awabboud 2014-09-30 13:49 Rev.: 52520

Add a new tuneable parameter for stability in CQMOM, expand cout messgae for CQMOM clipping.

36 lines of code changed in 6 files:

u0407585 2014-09-30 13:25 Rev.: 52519

added extra coefficients to Richards/fletcher devol model for derek

13 lines of code changed in 2 files:

cam 2014-09-30 08:52 Rev.: 52516

update --with-pidx to support latest version

7 lines of code changed in 2 files:

tsaad 2014-09-30 08:50 Rev.: 52515

memory leak cleanup

5 lines of code changed in 3 files:

dav 2014-09-29 14:44 Rev.: 52514


M configure
M configure.ac

- Give a more informative message (ie, which check is being called) when
checking for the fortran compiler.
- Remove the trailing / (if any) from boost include. This allows us to
remove /usr/include (or /usr/include/) if it is specified, as these
are compiler defaults and should not be used in a -I flag.

M StandAlone/sus.cc

- Remove duplicate "extern Mutex cerrLock;" statement.

114 lines of code changed in 3 files:

tsaad 2014-09-29 12:43 Rev.: 52512

add a test for interior boundary conditions. cleanup a few cout statements.

4 lines of code changed in 5 files:

tsaad 2014-09-29 12:07 Rev.: 52511

Add support for specifying interior boundaries through the Uintah interface. A great use of this is to attach surface boundaries to intrusions.
usage:
* To specify an interior boundary through the input file, use the <InteriorFace> xml block. This is identical to the <Face> specification used for domain boundaries. Example:
<InteriorFace ellipse="x-" origin="0.2 0.5 0.5" major_radius="0.2" minor_radius="0.05" angle="270" name="my boundary">
<BCType var="Dirichlet" label="f" value="10.0"/>
</InteriorFace>
* The minus/plus face spec will determine which cell iterator is returned. For a minus spec, the iterator will return the cells that are to the minus side of the interior boundary. The opposite is true for a plus spec
* TO access the cell iterator, use: Patch::getInteriorBndBCDataArray(Patch::FaceType face). All other access methods used domain boundaries remain the same. See Wasatch::BCHelper.cc for a great example of how to access the BC data associated with an interior boundary
* Interior boundaries currently support all boundaries supported by the existing interface (except for a side boundary - use rectangle instead): rectangle, circle, annulus, ellipse.

Limitations:
* Does NOT handle unions/differences. While support for those can be added later, there is no clear way on how to combine boundaries with different boundary conditions. Note that this limitation also applies to the current domain boundary conditions.
* Interior boundaries MUST be parallel to domain boundaries. For example, you cannot specify an inclined interior boundary.

Algorithmic details:
For a given grid resolution, the interior boundary is moved to the closest face (or layer of nodes parallel to that boundary). If the interior
boundary coincides with a cell center, then it is moved to the face side (minus/plus) that is
specified through the input file. The face side (minus/plus) determines which cell iterator is
returned. For a minus boundary, the cells on the minus side are returned. For a plus boundary,
the cells on the plus side are returned.

3248 lines of code changed in 36 files:

ahumphrey 2014-09-29 12:03 Rev.: 52510

Report #threads per MPI process only when using MPI_THREAD_MULTIPLE (level 3).

0 lines of code changed in 2 files:

u0407585 2014-09-28 08:29 Rev.: 52507

added 2-step richards-fletcher devol model, and fixed kobayashi-sarovim devol model. No regression tests should break

782 lines of code changed in 8 files:

mhomel 2014-09-26 18:26 Rev.: 52506

Added code that will hopefully prevent the code from hanging in the nonHardening return loop.

71 lines of code changed in 6 files:

jthornoc 2014-09-26 16:08 Rev.: 52505

Turning code back on that was causing a threaded rmcrt_DO test failure.

Attempts to reproduce the error with the latest code were unsuccessful,
including on the RT box and Todds baja box. Lets see what the RT says
tonight.

6 lines of code changed in 1 file:

jthornoc 2014-09-26 16:08 Rev.: 52504

More transport cleanup, this time with the scalars. This should not change answers.

32 lines of code changed in 1 file:

jthornoc 2014-09-26 16:08 Rev.: 52503

This commit removes some unused or otherwise unnecessary code in the
DQMOM transport. Answers should not change.

57 lines of code changed in 11 files:

derekhar 2014-09-26 09:39 Rev.: 52500

Added comments to improve clairity of coal drag model.

3 lines of code changed in 2 files:

jsutherland 2014-09-25 16:23 Rev.: 52499

Use SpatialOps::apply_pointwise() rather than old iterator loops for TabProps interface.
This should also work in thread-parallel, but not for GPU.

0 lines of code changed in 2 files:

jthornoc 2014-09-25 10:41 Rev.: 52497

Turned these local RT threaded tests off by accident in a previous commit.

Turning them back on.

2 lines of code changed in 1 file:

jsutherland 2014-09-25 08:57 Rev.: 52496

Get RMCRT compiling on GPU again.

1 lines of code changed in 1 file:

jthornoc 2014-09-24 12:01 Rev.: 52495

Turn off more mask features. Something isnt thread safe here.

5 lines of code changed in 1 file:

harman 2014-09-23 16:11 Rev.: 52493

When deleting radiometerPatchSet use removeReference()

9 lines of code changed in 2 files:

jthornoc 2014-09-23 13:48 Rev.: 52490

Fix a naming convention bug for DQMOM weights.

2 lines of code changed in 1 file:

jthornoc 2014-09-23 13:35 Rev.: 52489

Apply a BC after initialization for HL.
-Ben, Derek and Jeremy

0 lines of code changed in 2 files:

jthornoc 2014-09-23 13:35 Rev.: 52488

Do not clamp velocities in the RK averaging procedure.

Instead, use the volume fraction to avoid the averaging. BC application
upstream should have set these velocities appropriately.

10 lines of code changed in 1 file:

harman 2014-09-23 13:06 Rev.: 52487


Added a task to compute velTau_CC which is a modified
velocity used only to compute the viscous shear stress terms.
velTau_CC == vel_CC in the interior of the domain, in the extraCells
it's computed using:

// velTau_CC[ec] = (2 * vel_CC(ec) - vel_CC(interior) );

This fixes the bug of tau_XY != tau_YX at the domain walls.

0 lines of code changed in 12 files:

jthornoc 2014-09-23 10:22 Rev.: 52486

Removed unused constructor.

0 lines of code changed in 1 file:

jthornoc 2014-09-23 10:22 Rev.: 52485

DQMOM cleanup: Will change answers.
Changes are due to name changes.

Changed the naming convention on the internal coordinate variable (icv)
to only have the internal coordinate root, then append an _# for the
environment.

Moved the DQMOM eqn registration into the DQMOM factory.

212 lines of code changed in 12 files:

guilkey 2014-09-22 10:35 Rev.: 52483

Further minor optimization for cpdi when particle size is being scaled back.

0 lines of code changed in 2 files:

harman 2014-09-22 09:19 Rev.: 52478

Moved the creation of radiometerPatchSet outside of getPatches(). This will fix a memory leak.

24 lines of code changed in 3 files:

derekhar 2014-09-19 17:11 Rev.: 52476

Changed the Swirl interface. Swirl now uses the face-object's origin by default, unless specified otherwise in inputfile.

13 lines of code changed in 1 file:

ahumphrey 2014-09-19 15:22 Rev.: 52475

Quick refactor on previous Spatial task scheduling commit.

3 lines of code changed in 1 file:

ahumphrey 2014-09-19 15:21 Rev.: 52474

* Add support for handling "Spatially scheduled" tasks in the taskgraph checksum ( MPIScheduler::verifyChecksum() ).
An example would be the radiometer task, which is not scheduled on all patches (only a smalll subset).
* This commit begins to address making this checksum more sophisticated.

52 lines of code changed in 4 files:

guilkey 2014-09-19 12:11 Rev.: 52472

Fixing an error in yesterday's commit. The RT will still fail, but now the differences are
round off, not factors of 2.

45 lines of code changed in 1 file:

mhomel 2014-09-19 12:02 Rev.: 52470

Turned fastapproximatefunctions off by default. Can be enabled by #define MHfastfcns in Arenisca3.cc

30 lines of code changed in 4 files:

guilkey 2014-09-19 11:34 Rev.: 52469

Fixing my fixing of the warnings. I don't need what the allocate command returns, but I
still need to call the allocate command.

7 lines of code changed in 3 files:

harman 2014-09-19 11:15 Rev.: 52468

compiler warnings: removed unused variables

3 lines of code changed in 21 files:

guilkey 2014-09-19 11:15 Rev.: 52467

Silence some warnings so that Todd will crawl just a LITTLE bit out of my hiny.
Note: Remaining warnings under the MPM component, when compiled with gcc-4.7,
are not clear enough for me to fix at this time. All but 3 come from one file,
that is an imported library. The remainders are due to PETSc versionitis, or
are fortran related.

9 lines of code changed in 6 files:

ahumphrey 2014-09-19 08:20 Rev.: 52466

Remove errant call to unlock DataWarehouse level lock in clear(). This was benign but was confusing as it was never locked in the first place. Currently the level lock is not even being used.

2 lines of code changed in 1 file:

ahumphrey 2014-09-18 16:26 Rev.: 52464

Remove unnecessary nested locks that were causing EDEADLK in transferFrom() for ParticleVariables. These write locks are not recursive. Everything transferFrom calls already has the appropriate read and write locks. NOTE: the removed comment was stale and inaccurate.

3 lines of code changed in 1 file:

harman 2014-09-18 16:08 Rev.: 52463

This belongs with previous commit. Added d_ to all_patches and each_patch.

2 lines of code changed in 1 file:

harman 2014-09-18 13:15 Rev.: 52461

formatting
- add d_ to global variables eachPatch and allPatches.
- added vertical and white spaces everywhere.
- added {} to conditional and loops
- added method/function dividers
- no longer line wrapping at 80 columns, our screens are bigger than a dec terminal.
- 2 space indentation, not 3!

470 lines of code changed in 2 files:

guilkey 2014-09-18 12:46 Rev.: 52460

Only carry out re-calculations of dsize and relative_node_location if a particle actually
needs to be rescaled.

120 lines of code changed in 1 file:

harman 2014-09-18 12:18 Rev.: 52458

Added spatial scheduling machinery. It works with the OnTheFlyAnalysis:radiometer()
and needs to be incorporated into the RMCRT::RayTrace() task. It doesnt't work
with dbg builds due to an issue with schedule:verifyChecksum().


170 lines of code changed in 5 files:

harman 2014-09-18 10:03 Rev.: 52457

Don't exit out of script if one test fails go complete.

2 lines of code changed in 1 file:

mhomel 2014-09-17 17:12 Rev.: 52456

Fixed the poisson scaling bug, previously fixed in Arenisca 3
-This line, and those below, will be ignored--

M ConstitutiveModel/Arenisca4.cc

1 lines of code changed in 1 file:

mhomel 2014-09-17 16:33 Rev.: 52455

Fixed a bug in the step divide code, causing it to erroneously use an overly conservative dtep division.
New version is much faster and seems stable but let me know if you have any robustness issues.

51 lines of code changed in 2 files:

u0407585 2014-09-17 10:38 Rev.: 52453

Pulled user options from the coal source code, and made them required inputs in arche spec

255 lines of code changed in 20 files:

harman 2014-09-17 09:38 Rev.: 52451

Global:
formatting: added dividers between methods/functions

verifyChecksum()
- more descriptive error message
- add {}

111 lines of code changed in 2 files:

tsaad 2014-09-16 13:08 Rev.: 52450

Allow momentum equations to use flux limiters for momentum convective fluxes. Have not explored this yet in the presence of boundary conditions on momentum.

81 lines of code changed in 7 files:

jthornoc 2014-09-16 08:48 Rev.: 52449

Turning off mask storage for now.

1 lines of code changed in 1 file:

harman 2014-09-15 15:06 Rev.: 52448

Output to std:out the number of threads per mpi process.

1 lines of code changed in 1 file:

harman 2014-09-15 12:33 Rev.: 52447

doRecompileTaskgraph()
- recompile TG after a rad calc timestep to turn off all-to-all communication.

6 lines of code changed in 1 file:

derekhar 2014-09-15 11:13 Rev.: 52446

Fixed bug for the 2-D table coal interface. This correctly sets the Heat loss bounds for a 2-D table.

0 lines of code changed in 2 files:

harman 2014-09-15 09:13 Rev.: 52445

ScheduleTimeAdvance()
- added 1 level approach. Technically this isn't necessary but to be consistent with Arches/wasatch
and help keep me sane.

26 lines of code changed in 2 files:

mhomel 2014-09-14 21:39 Rev.: 52443

Added Lode angle dependence with Mohr Coulomb and Gudehus options. Added inputs for Lode angle functions and max princ stress.
Replaced some functions with lookup tables but still slow.

128 lines of code changed in 2 files:

mhomel 2014-09-14 20:13 Rev.: 52442

changed input types for Arenisca4 3rd invariant dependence paramters

110 lines of code changed in 1 file:

mhomel 2014-09-14 19:34 Rev.: 52440

Arenisca 4 no working with structure to support 3rd invariant dependence. Passes the A3 benchmarks.
Still needs inputs for Lode angle function and faster implementation.

205 lines of code changed in 2 files:

harman 2014-09-14 12:19 Rev.: 52438

ARCHES.py
added methane_rmcrt.ups to RMCRT section

methane_rmcrt.ups
- cleaned up and no longer computing enthalpy_do since it's not used.

methane_VR.ups
- input file that uses DataAnalysis:radiometer module.

1 lines of code changed in 1 file:

harman 2014-09-14 09:41 Rev.: 52437

removed semi-redundant ups file that doesn't run.

0 lines of code changed in 1 file:

harman 2014-09-13 12:25 Rev.: 52434

Added comments about expense of scheduling all-to-all variables

added
doRecompileTaskgraph() : can trigger a taskgraph recompile

Radiometer;
- added a conditional around the requires for the all-to-all variables.

The radiometer and ray tracing tasks should only require all-to-all variables on
calculation timesteps

67 lines of code changed in 5 files:

harman 2014-09-13 11:19 Rev.: 52433

Removed
-needRecompile()
-doCarryForward()

This functionality is provided from either RMCRTCommon or sharedState

1 lines of code changed in 2 files:

jthornoc 2014-09-12 16:57 Rev.: 52432

Add a transform function for f and HL.

109 lines of code changed in 3 files:

jthornoc 2014-09-12 16:57 Rev.: 52431

Add an optional boundary condition task to ArchesTaskInterface.

update all derived tasks.

359 lines of code changed in 32 files:

jthornoc 2014-09-12 16:57 Rev.: 52430

Name the boundary if no name is present.

17 lines of code changed in 2 files:

harman 2014-09-12 15:10 Rev.: 52429

SimulationState:
- added bool d_recompileTaskGraph, getRecompileTaskGraph(), setRecompileTaskgraph().
This is used by code deep inside a component with no easy access to the existing
simComp->needRecompile().

AMRSimulationController
needRecompile()
- check the sharedState if a taskgraph recompile has been requested.

21 lines of code changed in 3 files:

awabboud 2014-09-12 14:40 Rev.: 52428

fix dragmodel for w velocity dimension

1 lines of code changed in 1 file:

harman 2014-09-12 14:19 Rev.: 52426

Code cleanup.
added {}, horizontal and vertical spaces.

174 lines of code changed in 1 file:

jthornoc 2014-09-12 09:16 Rev.: 52425

Turn off the throw for now.

1 lines of code changed in 1 file:

jthornoc 2014-09-11 21:43 Rev.: 52424

Fixing a compile error. Not sure how it survived the Mac compiler.

33 lines of code changed in 3 files:

harman 2014-09-11 16:34 Rev.: 52423

Only schedule the "requires" for the super expensive
- abskg,
- sigmaT4
- cellType
variables on calculation timesteps. Carry forward timesteps ignore them.
The driving component must have needsRecompile() implemented with appropriate logic. This is
a 80% solution to the problem.


RMCRT_Test:
- added needsRecompile() which calls doCarryForward().
- formatting
doCarryForward()
- no longer passing in timestep

Ray.cc
sched_rayTrace()
- added logic around the requires.
doCarryForward()
- no longer passing in timestep

Radiometer.cc/ RMCRTCommon
doCarryForward()
- no longer passing in timestep

This shouldn't change the answers.

218 lines of code changed in 6 files:

jthornoc 2014-09-11 16:15 Rev.: 52422

Overload the register_variable function to reduce unnecessary
specification for NEWDW, COMPUTES, Ghost = 0.

31 lines of code changed in 3 files:

jthornoc 2014-09-11 15:36 Rev.: 52421

Now building all SVol masks.

Also, fixed a bug when using massFlowInlet on an any + face.

30 lines of code changed in 2 files:

jthornoc 2014-09-11 15:36 Rev.: 52420

Now parsing BC information from input file and creating masks for svol
variables.

35 lines of code changed in 1 file:

jthornoc 2014-09-11 15:36 Rev.: 52419

More refinement on the mask storage:
1) Remove polymorphism from the base mask struct. Didn’t seem necessary given the template parameters.
2) Rename the base struct to ‘MaskContainer’.
3) Don’t pass a ‘model field’ to get the mask back. This isn’t needed.
4) Created a nested set of maps that store patch ID and boundary names.
So, given then patchID and the boundary name typed on field, you get back the appropriate mask.

41 lines of code changed in 3 files:

awabboud 2014-09-11 13:38 Rev.: 52418

various fixes to the drag model, change Re limit to lessen discontinuity

31 lines of code changed in 2 files:

guilkey 2014-09-10 19:40 Rev.: 52416

Add a semicolon to fix compile error.

0 lines of code changed in 2 files:

guilkey 2014-09-10 19:16 Rev.: 52415

Set the default melting temperature to a large number, rather than room temperature, so that
if someone wishes to allow melting, they have to ask for it, rather than getting it by surprise.
Also, issue a warning if a particle's temperature exceeds the melting temperature, but the <do_melting>
flag is not set to true. Currently ElasticPlasticHP does nonsensical things, and the user should be
warned.

9 lines of code changed in 3 files:

awabboud 2014-09-10 15:43 Rev.: 52414

add operator scheduel to restart to fix nebo in arches restarts

2 lines of code changed in 1 file:

awabboud 2014-09-10 15:39 Rev.: 52413

move varlabel lookup from initialize to setup

14 lines of code changed in 1 file:

awabboud 2014-09-10 10:58 Rev.: 52411

change source wrapper to use updtaed values on second time substep

20 lines of code changed in 1 file:

mhomel 2014-09-10 03:37 Rev.: 52410

Eliminated the requirement for monotonically increasing insertParticle() times

44 lines of code changed in 1 file:

mhomel 2014-09-10 03:33 Rev.: 52409

Fixed a typo in the poisson scaling for arenisca3. Rewrote Arenisca4 to allow for 3rd invariant dependence. It builds but has not been tested yet at all.

1618 lines of code changed in 4 files:

jsutherland 2014-09-09 13:22 Rev.: 52408

More GPU-izing of wasatch.

49 lines of code changed in 7 files:

derekhar 2014-09-09 09:17 Rev.: 52407

Added "reqiures" for variables needed in cases using the RadProps library.

22 lines of code changed in 1 file:

harman 2014-09-09 09:14 Rev.: 52405

problemSetup()
- added bulletproofing. Prevent the user from specifying a variable that
will be ignored. This falls outside the UPS spec. checking.

14 lines of code changed in 1 file:

awabboud 2014-09-08 15:59 Rev.: 52404

fix the task dependency, this source is always newdw

2 lines of code changed in 1 file:

derekhar 2014-09-08 15:53 Rev.: 52403

Switched order of operations in RadProperties so that the _calc object is populated first.

2 lines of code changed in 1 file:

awabboud 2014-09-08 14:37 Rev.: 52402

Add a more generalized momentum drag source, adjust the momentum solver to actually call problemSetup on sources, fix drag model term.

352 lines of code changed in 8 files:

derekhar 2014-09-08 13:45 Rev.: 52401

This commit has added a regression test and is expected to cause failure in the nightly RT on coal_RadPropsPlanck.ups!


Fixed the interface for computation of absorption coefficients using the RadProps library.
The RadProps library can use one of two methods to compute the particle absorption coefficients (Planck or Rossland).
The new R-Test is similar to the coal-channel case called coal_RadPropsPlanck.

This commit also removed a work-around (or hack) due to a boundary condition bug, which was fixed by Tony Saad.

117 lines of code changed in 6 files:

harman 2014-09-08 10:01 Rev.: 52400

Added bulletproofing to user inputs. The starting and ending points must differ by at least one cell.

10 lines of code changed in 1 file:

jsutherland 2014-09-05 18:48 Rev.: 52399

minor cleanup

0 lines of code changed in 6 files:

jthornoc 2014-09-05 14:43 Rev.: 52397

Doxygen comments.

10 lines of code changed in 1 file:

jthornoc 2014-09-05 14:36 Rev.: 52396

First pass at implementing storage for masks. Still needs work.

196 lines of code changed in 5 files:

tsaad 2014-09-05 14:35 Rev.: 52395

forgot to commit this

2 lines of code changed in 1 file:

tsaad 2014-09-05 14:28 Rev.: 52392

Fix RMCRT on Wasatch's side to reflect recent changes upstream

15 lines of code changed in 3 files:

harman 2014-09-05 13:53 Rev.: 52390

passing grid into RMCRT problemSetup().


1 lines of code changed in 1 file:

harman 2014-09-05 13:26 Rev.: 52389

Now passing the grid into the problemSetup() tasks.

Added bulletproofing to VRLocationsMin/Max. They can't lie outside the
computational domain.

35 lines of code changed in 5 files:

harman 2014-09-05 08:08 Rev.: 52386

Removed function faceTypeToString(). It's redundant with getFaceName()

2 lines of code changed in 1 file:

awabboud 2014-09-04 14:42 Rev.: 52385

Remove the fluid phase model, this is handled by dragmodel.h now

0 lines of code changed in 2 files:

awabboud 2014-09-04 14:41 Rev.: 52384

Add the gas source term calcuation to the particle drag model, and rename/fix some of the variables.

82 lines of code changed in 2 files:

jthornoc 2014-09-03 10:02 Rev.: 52383

Missed registering some transport labels.

4 lines of code changed in 1 file:

awabboud 2014-09-03 09:35 Rev.: 52382

fix two of the new particle source terms

21 lines of code changed in 3 files:

awabboud 2014-09-03 08:53 Rev.: 52381

clean up unused area fraction from cqmom convection

23 lines of code changed in 2 files:

mhomel 2014-09-03 01:07 Rev.: 52380

added pressure dependent Poisson scaling, and cleaned up some old ifdef

135 lines of code changed in 2 files:

tsaad 2014-09-02 16:22 Rev.: 52379

First phase of improving the GeometryPiece et al. capabilities:
1. Add support for storing and returning a vector of points that are inside a named geometryPiece on a given patch
2. Add support for storing and returning a vector of points that are inside ALL named geometries for a given patch

usage:

To get the vector of points for a given geometry on a given patch, use:
const std::vector<SCIRun::Point>& insidePoints = GeometryPieceFactory::getInsidePoints(geomName, patch);

To get the vector of points for ALL named geometries on a given patch, use:
const std::vector<SCIRun::Point>& insidePoints = GeometryPieceFactory::getInsidePoints(patch);

Points are calculated once and then referenced every time the above calls are made. If a patch doesn't have any points inside the geometry then the above will return an empty vector. Note that these are static function calls so there is no need to carry around the list of geometry objects.

194 lines of code changed in 7 files:

August 2014 »

Generated by StatSVN 0.7.0