Also replace many instances of GetElementTransformation(0) with
GetTypicalElementTransformation() and instances of GetElementGeometry(0) with
GetTypicalElementGeometry().
Remove some "early returns" in PA code with empty partitions. The reasoning is
that the early returns create divergent code paths among different MPI ranks,
which can cause deadlocks if e.g. collectives are used in code paths on not all
ranks.
* AIR solver
* AIR: block-diag scaling, set AIR options
* new example
* SetAIROptions
* fix bugs
* more bug fix
* initialize solution vector in ode
* added blocksize
* Added command0line flags for AIR parameters.
* implemented ParCSRMatrix add in hypre and added MFEM interface.
no assumptions on "the same col_map_offd" as in the old "Add".
Should replace Add (no longer needed) with the new one
* Added absolute tolerance to HypreGMRES.
* Updated advection driver, added options to use BoomerAMG w/o gmres.
* Missing &
* AIR options: type of dropping entries for filtering A
* boomeramg print options.
* add a different \theta for R
* minor change
* for hypre 2.14
* test
* give hypre dof's coordinates
* nodal coarsening support
* changes according to the new changes in hypre/AIR [Sabs, BdiagInv]
* Added function to get CSR for on-processor rows of hyper matrix.
* Merged Ruipeng's edits with get processor function.
* Added hypre preconditioner for on-processor diret solves of triangular matrices.
* Added function to set max levels of BoomerAMG
* Added function to set max levels of BoomerAMG
* Merging with master
* Added extra functions to set specific BoomerAMG options such as interpolation, coarsening, etc.
* Added cycle option to boomeramg.
* Added funtionality to hypre boomeramg class
* Updated to support d1.5 AIR
* Added nAIR vs. lAIR options:
* Updated advection example
* Accidentally added build folder
* Small fixes to hypre files
* Add MFEM_adv steady-state-non-periodic test cases.
Add visualization.
Change config/defaults.mk to point to hypre dir.
* Corrected typo in MFEM_adv.cpp.
* Add multiple SN directions (azimuthal) as ImplicitSolve unknowns.
* Add higher azimuthal resolution test.
* Reduced number of AMG_solvers to one (not SN).
Clean up.
* First step towards the electron transport.
Hypre-mass-matrices-sparsity works.
* Code clean and tested.
Plasma profiles and all coeficients defined by functions.
* Transport example extended to run on ALE meshes (within the hydro code).
* Add include in exTRguts example.
* Source of particles inside Mult/ImplicitSolve.
* Scalar flux evolved within the TimeIntegrator.
* Integration over velocity range + a nice fig.
* AWBS/v3 model on, decelaretion integration on.
dfMdv source needs to be checked to give ne.
* AWBS working. Number of particles test ok.
* Both explicit and implicit integrators num of particles test ok.
* Add timer.
* Add AIR parameters parsing for AWBS runs.
* Add ne_ref and test nonlocal/local runs.
* Improved timer
Better matrix in memory treatment.
* Add velocity range input params.
* Add E field v-limiting
Lorentz Efield calculation.
* Add working scattering.
Low density -> positivity problems.
* A working setting (dv = - dt).
mpirun -n 4 exTRguts -m '../data/inline-quad.mesh' -rs 2 -vis -s 24 -ne 1e21 -sna 8 -vini 1.0 -vfin 0.0 -Ngr 100 -scit 1
Scattering is off.
* Scattering converges, yet unchecked.
mpirun -n 4 exTRguts -m '../data/inline-quad.mesh' -rs 1 -vis -s 24 -ne 1e21 -sna 4 -vini 1.0 -vfin 0.075 -Ngr 100 -scit 30
* Scaled -> nicer plot.
* Add exETRp example.
* Scattering is on.
* rm temp files, add user cmake to gitignore
* Added some more hypre options
* Updated example 9 to support implicit time stepping and AIR
* Added ex23 DG adv-diff with AIR for implicit solves
* example updates
* Fixing merge issue
* added two air options
* Updated AIR flags to be consistent with hypre master branch
* Removed include folder from repo
* removed build dir accidentally added..
* Removed TRT deve examples
* Removed TRT dev file
* Removed 23, MFEM_adv, updated ex9p to have AIR support
* Fixing travis stuff
* More travis
* Update hypre version in yml file for Travis
* Update hypre version in appveyor yml
* Update .appveyor.yml
* \ for windows?
* Added general DG blocksize, added extrat submatrix routine from hypre
* Fixed pointer
* Merge w/ master, fix one space in style
* Missed hypre dirs in appveyor and travis
* Reset appveyor/travis, added flag for hypre >= 21800.
* missed some hypre version #s
* Ex9 was modified in master, auto merge didn't work
* Ex9 running on quartz
* Affressing most of Dylan's comments
* Cleaning up hypre/AIR interface, addressing comments
* Some more comments
* Simplify AIR with ex9p
* Minor changes to hypre/AIR interface
* Use MPI_Session in ex9p
* Use Array instead of pointer
* Use HYPRE_Int or HYPRE_BigInt instead of int
* make style
* Minor
* Fix signed/unsigned int warning
* Removed boomeramg mult, added doxygen for wraphypreparvector
* Use MFEM's hypre memory wrappers
* Remove HypreBoomerAMG::Mult from header
* Added AIR comments, #if for hypre version
* Make input parameter const
* Use 3 procs for sample run
* Re-add comment
* Address review comments
* Address issues in WrapHypreParCSRMatrix
* Fix unintended change
* Fix memory leak in HypreParMatrix::MergeDiagAndOffd
* Small doxygen updates.
* Always call HypreParMatrix::Destroy when wrapping new vector
* Update CHANGELOG
Co-authored-by: Ruipeng Li <li50@llnl.gov>
Co-authored-by: Ruipeng Li <li50@quartz2498.llnl.gov>
Co-authored-by: holec1 <holec1@llnl.gov>
Co-authored-by: milan_holec <homijan@gmail.com>
Co-authored-by: Tzanio <tzanio@llnl.gov>
Co-authored-by: Will Pazner <will.e.p@gmail.com>
Co-authored-by: Veselin Dobrev <dobrev@llnl.gov>
there are ghost shared faces using amr-quad.mesh on 3 processors.
In ex9p, added a sample run on amr-hex.mesh on 3 processors so
have a test case with ghost shared faces in 3D.
Sohail Reddy