OH::Loss::ReactionProcessor(): uncommented

works better when new particles are shifter both in the space and
velocity. In the example is based on the SWMF's test test14_1d. Here,
the interaction between the neutrals and solar wind is commented to
simplify the numerics. The issue with the particle number in the
rarefied region is fixed.

maximal weight are selected for splitting

1. when splitting a particle: the original and the new particles get
velocity shift such that the momentum is conserved
2. when particles are merged: (a) a particles with a minimum weight is
identified, and (b) that particles is merged with another particles that
is closest in the velocity space

modified - the list of cells in the velocity space are source such
thatparticles are removed from the velocity-space cells that have most
of the particles. That does not really heliped with removing the
numerical artifact in the splitting/merging algorithm.

It seems that the PIC::ParticleSplitting::Split::Scatter() in the
previous commit (08bf5349) works fine
when particle merging is disabled

deleted

VelocitySpaceTable has been deleted

1. It seems that
PIC::ParticleSplitting::Split::SplitWithVelocityShift(0.001,160,200)
works better than
PIC::ParticleSplitting::Split::SplitWithVelocityShift(0.001,160,200)

2. the test was done with input/test/oh.shocktube.input, and SWMF's
PARAM.in srcOH/PARAM.in

3. The code dies after iteration 27-29 (-np=10)

the memory access model had been changed

the particle state vector is in the frame moving with solar wind.
SEP::ParticleMover_Droge_2009_AJ() is adapted for that.

particle injection can be determined runtime

of the SEP model

PARAM.in parser (srcInterface/amps_param_interface.cpp) that sets
parameters of the SEP model

spectrun is defined with vector 'SamplingHeliocentricDistanceList' that
can be modified before initialization of the model

be changed runtime

1. parameters of the diffusion coefficent SEP::Diffusion::Jokopii1966AJ
can be modified runtime
2. the diffusion coefficient model used in the simulation is defined
with the pointer SEP::Diffusion::GetPitchAngleDiffusionCoefficient and
can be modified runtime

commit (10549d7a) the search in the
lists is removed. That complicated the algorithm, bur should accelerate
it. The function needs some cleanup and a lot of comments.
Tested for OH case oh.shocktube.input

particle for removing is redone. In the current implementation, a mesh
is constructed in the velocity space. Particles are mergeed when in the
same cell. If there is no particles to merge, the mesh is coarsened.
NEED: accelerate the algorithm, and add comments.
Tested for OH case oh.shocktube.input

algorithm for selecting particles to be deleted is the same as in
PIC::ParticleSplitting::Split::SplitWithVelocityShift(). The fucntion
produces reasonable results when used for SWMF's test14_1d.

SEP::ParticleMover__He_2019_AJL(): GetX()/GetV() are replaced with another implementation because of an issues of accessing a velocity vector by pointer.... Mistery...

selecting particles to be deleted is changed

GCRs in 3D

added: utility utility/AssembleDistributedOutputFiles.pl to assemble files cleated by meshAMRgeneric::OutputDistributedDataTECPLOT()

2.  PIC::Mover::Simple(): v=PIC::ParticleBuffer::GetV(ParticleData) and x=PIC::ParticleBuffer::GetX(ParticleData) have beed replaced with PIC::ParticleBuffer::GetV(v,ParticleData)+ PIC::ParticleBuffer::GetX(x,ParticleData)+PIC::ParticleBuffer::SetX(x,ParticleData), which seems fixed running of the OH application
3. added: utility utility/AssembleDistributedOutputFiles.pl to assemble files cleated by meshAMRgeneric::OutputDistributedDataTECPLOT()

1. few types are fixed
2. addd flag AssembleDistributedOutputFile. Theflag determines whether output files created in OutputDistributedDataTECPLOT() by individual ranks would be assembled in a single file.

function to increase/reduce number of model particles in a cell.

particle tracking is disabled

corrected

corrected