--------O--------O--A-------A-------55555--2222
-------O-O------O--A------A-A------5-------------2
------O---O-----O-A------A---A-----5-------------2
-----OO-OO----OA------AA--AA----5555----2222
----O-------O---O-A----A-------A---------5--2----
---O---------O--O--A--A---------A--55555--22222
----Arbitrary---Kinetic--Algorithm--------------

AKA

Arbitrary Kinetic Algorithm for collisionless plasma modeling.

stack: C++11, MPI, HDF5, python2-3

HOWTO

1. before 'make' need to set in makefile
   HDF5_PATH= path to hdf5 lib
   MPI_PATH= path to mpi lib
   PYTHON_INC= path to python include
   PYTHON_LIB= path to python lib

2. for running default example from src/input/Initializer.py
   mpirun -n 2 aka.exe

3. normally need to set input file path containing Initializer.py
   mpirun -n 2 aka.exe PATH/TO/PYTHON/INPUT/FILE

4. before running need to create output folder and set in Initializer.py

5. for visualization use python notebook in folder NOTEBOOK/

MPI and HDF5 installation

download openmpi v 4.0.5 from
download.open-mpi.org/release/open-mpi/v4.0/openmpi-4.0.5.tar.gz

extract in folder /TEMP/FLD4INSTALLATION/, go to the folder

run in terminal:

1. ./configure --prefix=/FLD2LIBS/openmpi CFLAGS=-m64 CXXFLAGS=-m64 FFLAGS=-m64 FCFLAGS=-m64
   // need to specify 64-bit compilation via additional flag that should be passed to ALL compilers
2. make
3. make install

download hdf5 v1.10.5 from
http://hdfgroup.org/package/hdf5-1-10-5-tar-gz/?wpdmdl=13571

run in terminal:

1. ./configure --prefix=/FLD2LIBS/hdf5 --enable-parallel CC=/FLD2LIBS/openmpi/bin/mpicc
2. make
3. make install

physics included:

• space and time quantities are normalized on:

  • ion inertia length d0 = c/omega_pi
  • inverse ion gyrofrequency 1/Omega_ci

• electro-magnetic fields are calculated on two staggered grid (G1 and G2)
  using predictor-corrector scheme (see EleMagManager.cpp)

  E = -[VixB] + [JxB]/n - divPe/n - eta J

  B' = -rotE

  J = rotB

• ions are described in a kinetic way,
  dynamics is solved using first order interpolation of em fields

• electrons are described in a fluid way by:

  • density (equals to ion density ne=ni=n)
  • bulk velocity Ve = Vi - J/n
  • six-component pressure tensor Pij

• six-component pressure tensor P is integrated in time
  using subcycling explicit scheme, for implicit one need to build with flag -DIMPLICIT_PRESSURE

  P' = - Ve.∇P - P∇.Ve - P.∇Ve - (P.∇Ve)^T - q/m [ PxB + (PxB)^T ]

  where Ve - electron flow velocity
  q - electron charge
  m - electron mass
  B - magnetic field

• laser effects are imitated by ablation operator including
  heat operator and particle creation operator
  (see LaserMockManager.cpp)

• ablation operator works in localized area, called focal spot

• heat operator provides electron pressure increasing in the focal spot

• particle creation operator sustains constant target density

FEATURES

1. AKA is 3D, parallel (MPI), multispecies code

2. BC type for em fields and hydro quantities: 1 - periodic (see GridManager.cpp) 0 - damping layer (see EleMagManager.cpp and ClosureManager.cpp)

3. BC type for particle properties (see BoundaryManager.cpp):
   1 - periodic 0 - outflow // reaching border particle leaves domain forever

4. for small scale dissipation use resistivity (eta) parameter

5. for pressure tensor integration need to set:

   • electron mass
   • relaxation factor for izotropization operator (see ClosureManager.cpp)
   • smooth stride for pressure tensor smoothing

6. use 'make FLAGS=-DLOG' to see some logs
   to set debug log level see Logger.hpp
   for extra logging use -DHEAVYLOG

7. for each particle type have to specify:

   • mass
   • charge
   • if particles are frozen in space (skip pusher phase)
   • particles per cell number

8. if number of laser focal spots is more than zero,
   auxiliary particle type is reserved for the loading fraction

TROUBLESHOUTING:

1. for OS X: ....python2.7/pyport.h:731:29:
   note: expanded from macro 'toupper'
   solution:
   add #ifndef __cplusplus... line 699 + #endif
   in /pyport.h line 722