December 2010 Commit Log

Number of Commits:
102
Number of Active Developers:
9
harman 2010-12-29 16:35 Rev.: 46650

interpolateToParticlesAndUpdate_CFI()
- use px_coarse from the old_dw when computing the shape function weights.

This eliminates the hash in pure translation plots!

0 lines of code changed in 2 files:

harman 2010-12-29 15:05 Rev.: 46649

debug_CFI()
-color the particles by the sum of the interpolation weights.

51 lines of code changed in 1 file:

luitjens 2010-12-28 18:10 Rev.: 46648

Added some error checking to the geometry object code. Requesting data that was not added to the object will now throw an error.

0 lines of code changed in 2 files:

luitjens 2010-12-28 17:48 Rev.: 46647

Added the ability to specify the refinement level with refinement objects. Use -1 to indicated the finest level.

7 lines of code changed in 1 file:

harman 2010-12-28 17:42 Rev.: 46646

errorEstimate()
- simplified logic for setting refinement flags. Just test the cell's lower
and upper point and determine if it's inside a geom_object.


8 lines of code changed in 1 file:

jthornoc 2010-12-28 17:32 Rev.: 46645

Initialize boolean.

0 lines of code changed in 2 files:

luitjens 2010-12-28 17:12 Rev.: 46644

Added the code for the Integer data type in geometry object fields.

8 lines of code changed in 3 files:

harman 2010-12-28 17:05 Rev.: 46643

problemSetup()
- added missing "res" to the GeometryObject::DataItem

2 lines of code changed in 1 file:

tsaad 2010-12-28 07:09 Rev.: 46642

Updated boundary condition treatment to infer BCEvaluator type from the boundary condition operator. This goes along with the recent update in SpatialOps.

0 lines of code changed in 2 files:

harman 2010-12-27 18:29 Rev.: 46641

changed getInitialData_double(velocity) -> getInitialData_Vector(velocity)

0 lines of code changed in 2 files:

harman 2010-12-27 17:22 Rev.: 46640

code generalization:

A GeometryObject can now have an arbitrary number of double,Vector,
IntVectors,Points variables associated with that object. The assumption that
each object will have a "res" and "velocity" has been removed. Justin & Todd

149 lines of code changed in 12 files:

harman 2010-12-23 17:42 Rev.: 46638


coarsenNodalData_CFI()
-only overwrite coarse level nodal data if there is fine level data gmass is
> d_SMALL_NUM_MPM

computeAndIntegrateAcceleration()
- compute gacceleration and gvelocity_star if gmass > d_min_mass_for_acceleration


Added debugging code all over the place.


71 lines of code changed in 1 file:

harman 2010-12-23 16:47 Rev.: 46636

sorted #includes

problemSetup()
- now reading geom_objects that define where the static refinement flags are layed down.

scheduleInitialErrorEstimate()
This just calls scheduleErrorEstimate()

ErrorEstimate()
- refinment flags are static for now and defined in the input file:
Usage:

<AMR>
<MPM>
<Refine_Regions>
<geom_object>
<box label = "regrid">
<min> [-0.2, -0.2,-0.001] </min>
<max> [ 0.2, 0.2, 0.001] </max>
</box>
<res> [2,2,1] </res>
<velocity> [0,0,0] </velocity> <!-- this sucks -Todd -->
</geom_object>
</Refine_Regions>
</MPM>
</AMR>

113 lines of code changed in 3 files:

harman 2010-12-23 16:32 Rev.: 46635

removed the diagnostic label gSumInterpWeights. Keeping track of the weights
isn't useful.

4 lines of code changed in 5 files:

harman 2010-12-23 11:35 Rev.: 46634

Remove deadwood
- findCellAndWeights() - AMR version
- findFinerNodes()

0 lines of code changed in 4 files:

guilkey 2010-12-22 13:19 Rev.: 46632

Uncomment some #undef's so that it isn't ambiguous.

0 lines of code changed in 2 files:

jsutherland 2010-12-21 20:23 Rev.: 46631

- Bug fix for parallel execution when we build trees unique per patch.
- Turn on unique trees per patch so Tony can hammer out BC issues.

0 lines of code changed in 14 files:

jsutherland 2010-12-21 20:22 Rev.: 46630

Activate unique expressions per patch.
Tony, have a go at boundary conditions now...

5 lines of code changed in 2 files:

guilkey 2010-12-21 18:56 Rev.: 46629

A slightly improved version of dumping out the UIs before and after KMMCHK.

2 lines of code changed in 1 file:

guilkey 2010-12-21 16:00 Rev.: 46627

Print out the user inputs before and after calls to KMMCHK, so the user can
see how Kayenta saw fit to modify the inputs.

12 lines of code changed in 1 file:

harman 2010-12-21 10:33 Rev.: 46626

added bulletproofing to inputs

0 lines of code changed in 2 files:

jsutherland 2010-12-17 19:03 Rev.: 46625

Cleanup after Amir's checkin.

- remove a bunch of temporary latex files that were committed.
- Fix some formatting (indentation)
- eliminate iterators in favor of newly added expression templates.

0 lines of code changed in 4 files:

amir 2010-12-17 18:41 Rev.: 46624

This is some notes about what I have changed in Wasatch in uintah

Here is a summary of the added files:
- uintah/src/CCA/Components/Wasatch/Expressions/MMS/TaylorVortex.h
- uintah/src/StandAlone/inputs/Wasatch/MMS_P.ups
- uintah/doc/Other/Wasatch/MMS/(All the containing documentary files over there - 17 files)

Here is a summary of the modified files:
- uintah/src/CCA/Components/Wasatch/Expressions/BasicExprBuilder.cc
- uintah/src/StandAlone/inputs/UPS_SPEC/wasatch_spec.xml
- uintah/src/CCA/Components/Wasatch/CoordHelper.cc


743 lines of code changed in 3 files:

jsutherland 2010-12-17 18:10 Rev.: 46623

First cut at per-patch graphs/expressions.

Right now I am duplicating the status-quo behavior, but using a new interface that will allow us to (hopefully) migrate easily over to per-patch graphs/expressions.

Basically, a TaskInterface now manages creation of ExpressionTree objects, and can manage these so that either one is built for all patches or one is built per patch. Currently, only the one-for-all situation has been tested. Soon we will move to one graph per patch. That will facilitate boundary condition implementation.

195 lines of code changed in 6 files:

jsutherland 2010-12-17 18:10 Rev.: 46622

Disallow expression creation from within the BC interface.

1 lines of code changed in 1 file:

jsutherland 2010-12-17 18:09 Rev.: 46621

Minor cleanup.

66 lines of code changed in 5 files:

jsutherland 2010-12-17 18:08 Rev.: 46620

Remove a garbage file that was accidentally committed.

0 lines of code changed in 1 file:

harman 2010-12-17 17:57 Rev.: 46619

Global:
replaced
#if 0 -> #ifdef DEBUG

computeInternalForce()
- initialize gstress

computeAndIntegrateAcceleration()
- variable name changes [c] -> [n]
- velocity -> gvelocity,
- added debugging code.

setGridBoundaryConditions()
- added conditional around setting BCs. Only set the BCs at the edge of the
computational domain, not at the CFI.

70 lines of code changed in 1 file:

harman 2010-12-16 17:23 Rev.: 46616

using d_nPaddingCells_coars=1 & NGP = 1

coarsenNodalData_CFI()
- zero gAcceleration and gVelocityStart at the CFI on the coarse level nodes.
This eliminates the double counting on the coarse nodes.

Added debugging code to identify bugs, currently turned off.

79 lines of code changed in 1 file:

luitjens 2010-12-16 15:55 Rev.: 46615

Add the executing processor to its neighorhood. Previously if a processor had no patches on all levels it would not be in its own neighborhood. This in turn caused once per proc tasks to cause a deadlock.

0 lines of code changed in 2 files:

harman 2010-12-15 16:35 Rev.: 46614

-add single title for the montage of plots
- minor formatting changes

17 lines of code changed in 1 file:

harman 2010-12-14 17:51 Rev.: 46613

changed assert to (s>=0)

1 lines of code changed in 1 file:

harman 2010-12-14 17:22 Rev.: 46612

The Lz+ conditional statement was using the wrong zone of influence component.
Added assert to catch when a shape function is negative.

5 lines of code changed in 1 file:

harman 2010-12-14 16:24 Rev.: 46611

Changed the gnuplot script to a csh script

usaage:

plot_dat_files <uda>


83 lines of code changed in 2 files:

harman 2010-12-14 13:07 Rev.: 46610

gnuplot script for plotting conserved quantities.

0 lines of code changed in 1 file:

jthornoc 2010-12-14 12:42 Rev.: 46609

Per Todd's recommendation, using a different method from isnan to try to
catch the random Hypre segfaults.

0 lines of code changed in 2 files:

jsutherland 2010-12-14 09:36 Rev.: 46608

use field expression templates in wasatch expressions

0 lines of code changed in 2 files:

jsutherland 2010-12-14 09:36 Rev.: 46607

Add some documentation.

0 lines of code changed in 2 files:

harman 2010-12-13 18:26 Rev.: 46606

added the global variable
d_nPaddingCells_coarse: Number of cells on the coarse level that contain particles and surround a fine patch.

Turned off debugging output.

48 lines of code changed in 2 files:

harman 2010-12-13 17:42 Rev.: 46605


getOtherLevelPatches()
use IntVector::doseIntersect instead of Box::overlaps() when
deciding if a patch is within requested index space.
The call Box::overlaps() fails to do the right thing when the
fine patch boundary coincides with the coarse patch boundary

This fixes a long standing bug when getting fine level patches in ICE::reflux
code and AMRMPM.cc

0 lines of code changed in 2 files:

harman 2010-12-13 16:40 Rev.: 46604

added missing %s in warning message

Search for the phrase "Caught exception" in sus's output instead of "exception"
The user could have the word "exception" in their code modifications which would
trigger a false positive.


0 lines of code changed in 2 files:

harman 2010-12-13 13:28 Rev.: 46603

interpolateParticleToGrid_CFI()
- use Patch::getOtherLevelPatches with a layer of padding cells instead of
Patch::getCoarseLevelPatches().

debug_CFI()
- use Patch::getOtherLevelPatches() with a layer of padding cells instead of
Patch::getFineLevelPatches()

We need the layer of padding cells in addition to the extraCells.

There is a bug in Patch:: getOtherLevelPatches that will be fixed soon.

12 lines of code changed in 1 file:

jthornoc 2010-12-13 10:40 Rev.: 46602

Added a check for nan's while we try to track down these random hypre
segfaults.

29 lines of code changed in 1 file:

jthornoc 2010-12-13 09:41 Rev.: 46601

Fixed some compiler warnings.

19 lines of code changed in 3 files:

jthornoc 2010-12-13 09:05 Rev.: 46600

Fixed a couple of warnings.

1 lines of code changed in 1 file:

jthornoc 2010-12-12 12:19 Rev.: 46599

Update to allow for TabProps generated tables with mixture fraction
variance. Mixing model can be turned on/off using the
<use_mixing_model> tag under <Properties>.

0 lines of code changed in 4 files:

tsaad 2010-12-10 15:52 Rev.: 46596

Switched to using pointers in passing the arguments for the BCHelper

9 lines of code changed in 3 files:

harman 2010-12-10 15:12 Rev.: 46595

added the task debug_CFI() as a debugging task.
- This task colors the particles that are retrieved from the coarse level along
the CFI. It uses the same logic as interpolateParticlesToGrid_CFI() to determine
the particle set. You can visualize the particles and see if the particle
set is correct.

139 lines of code changed in 2 files:

harman 2010-12-10 15:01 Rev.: 46594

scheduleParticleRelocation()
-added comment

0 lines of code changed in 2 files:

guilkey 2010-12-10 14:27 Rev.: 46593

Remove some spew

0 lines of code changed in 2 files:

harman 2010-12-10 11:44 Rev.: 46591

put(ParticleVariableBase)
- Improved the exception message.

0 lines of code changed in 2 files:

harman 2010-12-10 11:19 Rev.: 46590

Only set the variable startpath once.

0 lines of code changed in 1 file:

luitjens 2010-12-09 17:51 Rev.: 46589

Fixed a bug in my last commit.

8 lines of code changed in 1 file:

harman 2010-12-09 16:04 Rev.: 46588

Added uintah to all::

removed some dead wood.

0 lines of code changed in 2 files:

luitjens 2010-12-09 15:23 Rev.: 46587

Changed the way getParticleSubset() with an arbitrary range is queried. You now pass in the patch from which you want the data as the key and no longer pass in a level.

35 lines of code changed in 8 files:

harman 2010-12-09 12:13 Rev.: 46584

getCoarseLevelRange()
- add the truncation offset for each direction, x, y, z, independently.

0 lines of code changed in 2 files:

luitjens 2010-12-09 09:19 Rev.: 46583

Removed some unnecessary asserts

0 lines of code changed in 1 file:

jsutherland 2010-12-08 18:35 Rev.: 46581

Add another rtest and flip the switch that Justin put in to force variables to be allocated per patch rather than groups of patches (no memory windowing)

9 lines of code changed in 5 files:

harman 2010-12-08 18:23 Rev.: 46580

Global:
getCoarseLevelRange()
- now passing in bool returnExclusiveRange.
if set to true then add the offset to ch.

AMRMPM.cc
interpolateParticlesToGrid_CFI()
Working on getting proper number of coarse level particles in the
padding cells that surround the fine patch.

42 lines of code changed in 10 files:

harman 2010-12-08 17:49 Rev.: 46579

added more dividers

21 lines of code changed in 1 file:

harman 2010-12-08 17:48 Rev.: 46578

Formatting.
- Consolidated all getParticleSubsets calls together.
- tabs
- added dividers
- cleaned up parameter lists.

Slightly easier to read

449 lines of code changed in 1 file:

tsaad 2010-12-08 17:01 Rev.: 46577

fixed typo in BCSide

41 lines of code changed in 1 file:

harman 2010-12-08 15:22 Rev.: 46575

A present for Justin.

Added a helper script for local regression testing. A symbolic link is added
to the Standalone directory when:

make uintah

is executed. This helper script just executes:

make gold_standards
or
make runLocalRT

24 lines of code changed in 3 files:

harman 2010-12-08 14:41 Rev.: 46574

modified the replace gold standards message to include the path to the
replace_all_GS script.


9 lines of code changed in 1 file:

luitjens 2010-12-08 13:54 Rev.: 46573

Sanity check the variable ranges after applying ghost cells.

5 lines of code changed in 1 file:

luitjens 2010-12-08 11:27 Rev.: 46572

Added the ability to turn off variable combining on the OnDemandDataWarehouse. If disabled each patch/variable combination will get its own memory.

To disable combining set OnDemandDataWarehouse::d_combineMemory=false. This should only need to be done once as it is a static member of the class and is shared by all instances.

206 lines of code changed in 2 files:

harman 2010-12-08 11:25 Rev.: 46571

remove temporary files when done with them.

0 lines of code changed in 4 files:

harman 2010-12-08 11:23 Rev.: 46570

copy the text file
sus.log.txt, timestamp, restart_timestamp, malloc_stats, restart_malloc_stats
to the gold standard directory when it is initially created.

0 lines of code changed in 2 files:

guilkey 2010-12-07 19:44 Rev.: 46569

Use a better polar decomposition algorithm.

0 lines of code changed in 2 files:

guilkey 2010-12-07 19:40 Rev.: 46567

Change the way the surface normals are computed, borrowing on ideas of
John Nairn. As before, a normal is computed for every material at each node.
Then, the normal with the largest mass gradient is chosen as "the" normal,
and the normals of the other materials are assigned based on that. This
gives more reliable normals, reduces issues with corners, and, by making the
normals for all materials colinear, should help with momentum conservation.

0 lines of code changed in 2 files:

luitjens 2010-12-07 11:55 Rev.: 46566

Removed an unecessary and incorrect error checking.

0 lines of code changed in 1 file:

harman 2010-12-07 10:20 Rev.: 46565

More descriptive exception message

3 lines of code changed in 1 file:

dav 2010-12-06 20:23 Rev.: 46564

Committing configure to match previous commit of configure.ac

0 lines of code changed in 2 files:

dav 2010-12-06 20:23 Rev.: 46563

Updates for (Cray-Xt) Franklin computer... Not checking in 'configure' from this machine as autoconf is wrong version.

0 lines of code changed in 2 files:

jpeterson 2010-12-06 18:31 Rev.: 46562

Add bulletproofing for reasonable range for the Poisson ratio of the material.

0 lines of code changed in 2 files:

harman 2010-12-06 16:52 Rev.: 46559

global:
- added a few comments
- cleaned up debugging output

problemSetup()
- simplfied some confusing logic. Restarts now work

39 lines of code changed in 1 file:

harman 2010-12-06 13:31 Rev.: 46557

coarsenNodalData()
- rewrote task.
- only schedule this task on the coarse levels.
- You can't modify data on a different level. In the previous implementationn
I was looping over fine level patches trying to modify coarse data at the CFI.

104 lines of code changed in 1 file:

jpeterson 2010-12-06 12:25 Rev.: 46556

Added if clause to prevent crashing (sigsegv) when crack model not specified

0 lines of code changed in 2 files:

jthornoc 2010-12-06 11:53 Rev.: 46555

Added some doxygen comments. Also, resolved the spline information
outside of the cell loop for better efficiency in TabPropsInterface.cc.

48 lines of code changed in 4 files:

harman 2010-12-06 11:18 Rev.: 46554

added output indentation.

0 lines of code changed in 2 files:

jthornoc 2010-12-04 06:26 Rev.: 46553

Comment out debugging output for efficiency.

2 lines of code changed in 1 file:

tsaad 2010-12-04 01:34 Rev.: 46552

TSAAD: cleanup of boundary condition handler code.

51 lines of code changed in 1 file:

harman 2010-12-03 19:00 Rev.: 46551

MPMCommon
- added printTask with different input parameters
AMRMPM
scheduleInterpolateParticlesToGrid_CFI()
- requiring variables with gac and padding cells. This may get us
past the issue of getting coarse level particles in a region surrounding
the fine patch.

scheduleCoarsenNodalData_CFI()
- changes so it will work on > 2 levels

scheduleInterpolateToParticlesAndUpdate_CFI()
- added scheduling for gSumWeights.

56 lines of code changed in 5 files:

jsutherland 2010-12-03 18:15 Rev.: 46550

Fix a bug in calculating coordinate field values.

8 lines of code changed in 1 file:

tsaad 2010-12-03 16:48 Rev.: 46549

TSAAD:
1. Added boundary condition treatment for Wasatch. This required modifying the equation adaptors to store the xml input information about each trasnport equation.
2. Added a test case for boundary conditions.
NOTE: due to issues with extracells and boundarylayer cells, this code is NOT functional yet.

689 lines of code changed in 8 files:

harman 2010-12-03 16:06 Rev.: 46548

removed debugging output

0 lines of code changed in 1 file:

harman 2010-12-03 15:26 Rev.: 46547

Changes so the user can have "-" in their filenames.

Implemented a different way of extracting the log comments from the input
arguments. You need to take care when the user has " " in their comments.
This worked in the past but for some reason today it didn't.

28 lines of code changed in 5 files:

harman 2010-12-03 11:56 Rev.: 46545

fixed conditional statement syntax error.

0 lines of code changed in 2 files:

jthornoc 2010-12-03 11:05 Rev.: 46544

TabPropsInterface:
Removed a variable that wasn't needed and used the private class
variable instead.

ExplicitSolver.cc:
Rearranged the dummy solve a little. This shouldn't change the answers
but makes more intuitive sense (expect the dummy solve makes no sense --
so just try to figure that one out)

ScalarEqn.cc:
In addition to the clipping after the averaging, do clipping after the
FE update to ensure bounded scalars.

44 lines of code changed in 5 files:

jsutherland 2010-12-02 20:51 Rev.: 46541

Missed addition of a file previously.

213 lines of code changed in 1 file:

jsutherland 2010-12-02 20:34 Rev.: 46540

remove "inline" keyword from a few methods that are now in the UpwindInterpolant.cc file.

0 lines of code changed in 2 files:

jsutherland 2010-12-02 20:31 Rev.: 46539

Update TabPropsEvaluator due to recent TabProps changes.

0 lines of code changed in 2 files:

jsutherland 2010-12-02 20:23 Rev.: 46538

Move a bunch of stuff from header files to .cc files, and use explicit template instantiation.

This should help speed up compilation times and may reduce code bloat a bit.

1165 lines of code changed in 23 files:

luitjens 2010-12-02 16:27 Rev.: 46535

Removed extraCells as a static member of Patch. This member didn't make sense since extraCells are specified per level not globally.

39 lines of code changed in 6 files:

harman 2010-12-02 15:38 Rev.: 46534

removed darwin local test

0 lines of code changed in 2 files:

harman 2010-12-02 13:56 Rev.: 46533

Replaced calls to getExtraHigh/LowIndex with the equivalent SFC(x,y,z) calls.

These calls were intended to be used with boundary layer cells, not as
a way to offset the iterators.

This will not change the answers.

0 lines of code changed in 6 files:

luitjens 2010-12-02 12:00 Rev.: 46531

Updated the computeExtents query to use other queries in the Patch class in order to ensure consistency.

29 lines of code changed in 2 files:

jpeterson 2010-12-01 16:54 Rev.: 46530

Last commit caused an error if useCrackModel was turned to false, being unable to compute the crackedEnough CC variable. This commit fixes that and consolidates a few chunks of code.

19 lines of code changed in 1 file:

harman 2010-12-01 16:32 Rev.: 46529

Removed dead code.

0 lines of code changed in 4 files:

harman 2010-12-01 16:30 Rev.: 46528

I can't remember is it make cleanreally or make reallyclean. Now either will work.

3 lines of code changed in 1 file:

harman 2010-12-01 15:40 Rev.: 46527

Global: formatting


Added new task Nodal_velocity_temperature()
gVelocity /= gmass
gTemperature /= gmass
This task is scheduled after the fine level nodal data has been coarsened.
Removed this computation from InterpolateParticlesToGrid


InterpolateParticlesToGrid_CFI()
- added the missing variables that needed to be interpolated.

CoarsenNodalData_CFI()
- added the missing variables

Serial, 2D, 2 level, refinement ratio 1, pure translation, seems to work.

341 lines of code changed in 2 files:

luitjens 2010-12-01 13:10 Rev.: 46526

Added some comments

0 lines of code changed in 2 files:

November 2010 »

Generated by StatSVN 0.4.0