BoundaryConditions
(28 files, 4444 lines)
PatchBVH
(9 files, 882 lines)
Variables
(81 files, 12020 lines)
[root]/Core/Grid
BoundaryConditions
(28 files, 4444 lines)
PatchBVH
(9 files, 882 lines)
Variables
(81 files, 12020 lines)
| Author | Changes | Lines of Code | Lines per Change |
|---|---|---|---|
| Totals | 161 (100.0%) | 564 (100.0%) | 3.5 |
| harman | 12 (7.5%) | 228 (40.4%) | 19.0 |
| dav | 3 (1.9%) | 154 (27.3%) | 51.3 |
| tsaad | 5 (3.1%) | 62 (11.0%) | 12.4 |
| jas | 109 (67.7%) | 59 (10.5%) | 0.5 |
| jsutherland | 23 (14.3%) | 52 (9.2%) | 2.2 |
| qymeng | 9 (5.6%) | 9 (1.6%) | 1.0 |
Finalize code for using Wasatch in Arches. The following changes have been made:
1. Add one way coupling between Arches fluid dynamics and Wasatch transport equations.
2. Add boundary conditions to wasatch-in-arches. All models used from Wasatch have their BCs set in Wasatch.
3. Add a Wasatch material to get things ready for MPMARCHES.
4. Wasatch in Arches now works with multipatch, multiprocessor with mpi (multipatch per processor also).
5. Multistage integrator for wasatch-arches using the Arches RK averaging interface.
6. Modify wasatch transport equations to automatically fill in the root IDs for the solution-graph-helper.
7. Add wasatch-arches input file test. This test is based on the Arches helium_1m.ups.
If you want to test Wasatch-in-Arches, there are two pieces of code in Wasatch that you must comment. This decision is based on our preference to use GhostCells instead of ExtraCells in Wasatch. When these lines of code are commented, wasatch will play well with extraCells. Please contact me or James Sutherland for additional information on what lines must be commented.
61 lines of code changed in 4 files:
Output task internal thread waiting time.
2 lines of code changed in 2 files:
Protect level cache.
0 lines of code changed in 4 files:
Spam diet!
Show the warning "No BoundaryConditions specified \nThis message will only occur once"
once, not once for every processor.
5 lines of code changed in 1 file:
MPMARCHES:scheduleMomExchange()
- only compute d_cellInfoLabel for the arches matl.
PressureSolverV2.cc
- initialize d_indx = -9
- addHydrostaticTermtoPressure()
use locally defined indx instead of d_indx. This task is called by
MPMARCHES and d_indx isn't defined.
UnknownVariable.cc
- removed conditional around adding matl index to exception message.
4 lines of code changed in 2 files:
Update copyright date.
4 lines of code changed in 8 files:
added another instance of printTask()
0 lines of code changed in 4 files:
Update copyright year.
53 lines of code changed in 99 files:
problemSetup()
- added comment and proc0cout statement.
7 lines of code changed in 1 file:
Fixes for static compilation for gcc 4.6.
1 lines of code changed in 1 file:
rip out the obselete "SpatialOps" component.
0 lines of code changed in 4 files:
Eliminate "using namespace std" and similar from header files. That is a big no-no!
16 lines of code changed in 10 files:
Additional documentation for getNumCells.
1 lines of code changed in 1 file:
add a ALL_CELLS value for requesting all cells in a level
2 lines of code changed in 1 file:
Support more than 2 ghost cells.
This will allow a task to request any number of ghost cells up to the whole grid.
5 lines of code changed in 2 files:
IntVector.h
- added roundNearest(). This will round the Vector v to the nearest integer
Level.cc
findNodeIndexRange(), findCellIndexRange(), findInteriorCellIndexRange()
findInteriorNodeIndexRange()
using roundNearest() on the lowIndex and highIndex.
This change fixes the following bug.
----------------------------------------------------------------
WORKS
<Box label="1">
<lower> [0.00,-0.0, 0.0] </lower>
<upper> [2.4, 0.1, 0.004] </upper>
<extraCells> [1,1,1] </extraCells>
<patches> [8,1,1] </patches>
<resolution> [160,12,1] </resolution>
</Box>
Debugging output:
findCellIndexRange: l [-1.0000000000000000e+00 -1.0000000000000000e+00 -1.0000000000000000e+00]
h [1.6100000000000000e+02 1.3000000000000000e+01 2.0000000000000000e+00]
lowIndex [int -1, -1, -1] highIndex [int 161, 13, 2]
----------------------------------------------------------------
NOWORK!
<Box label="1">
<lower> [0.00,-0.001, 0.0] </lower>
^^^^^^ <ONLY DIFFERENCE> <<<<<<<<<<<<<<<<<<<<<<<<,
<upper> [2.4, 0.1, 0.004] </upper>
<extraCells> [1,1,1] </extraCells>
<patches> [8,1,1] </patches>
<resolution> [160,12,1] </resolution>
</Box>
findCellIndexRange: l [-1.0000000000000000e+00 -9.9999999999999978e-01 -1.0000000000000000e+00]
^^^^^^^^^^^^^^^^^^^^^^^???????
h [1.6100000000000000e+02 1.3000000000000000e+01 2.0000000000000000e+00]
lowIndex [int -1, 0, -1] highIndex [int 161, 13, 2]
^^ <<<<<<<<<<<<<This cause all sorts of problems downstream.
----------------------------------------------------------------
**** This should not change the answers *****
11 lines of code changed in 1 file:
Moved
coarsenDriver_std() and coarsenDriver_massWeighted ()
from MPMICE to AMR_CoarsenRefine.cc, so other tasks can use them.
201 lines of code changed in 3 files:
return LevelP by reference rather than by value (is there a problem with the Handle class copy semantics?)
4 lines of code changed in 3 files:
- eliminate constLevelP typedef as it was not used and confused me.
- add a few convenience functions to extract LevelP rather than Level* from patches.
20 lines of code changed in 4 files:
Fixed circular dependencies recently added to the code due to code
consolidation.
1) Consolidated CCA/Components/Arches and Arches/Radiation into a single library
instead of two separate libraries... There isn't any good reason
that I know of to keep them separate, and this allows them to
'share' some common routines w/o creating a circular dependency.
(There are still a couple of additional Arches sub-shared
libs... (fortran/Mixing). These could easily be consolidated too if someone
wanted it done.)
Note, previously, in CCA/Components/sub.mk, many of the Arches sub-shared libs
where specifically specified. I've removed these and moved them
directly into the CCA/Components/Arches/sub.mk file to make it a
little easier to follow what is going on.
3) Moved DbgOutput out of Core/IO and into Core/Grid as it is related to
the Grid classes... Also, Core/Grid depends on Core/IO, so having
Core/IO stuff depend on Core/Grid causes a circular dependency.
154 lines of code changed in 3 files:
(2 more)