.----------------.  .----------------.  .----------------.  .----------------.
| .--------------. || .--------------. || .--------------. || .--------------. |
| |  ________    | || |     _____    | || |     ______   | || |  _________   | |
| | |_   ___ `.  | || |    |_   _|   | || |   .' ___  |  | || | |_   ___  |  | |
| |   | |   `. \ | || |      | |     | || |  / .'   \_|  | || |   | |_  \_|  | |
| |   | |    | | | || |      | |     | || |  | |         | || |   |  _|  _   | |
| |  _| |___.' / | || |     _| |_    | || |  \ `.___.'\  | || |  _| |___/ |  | |
| | |________.'  | || |    |_____|   | || |   `._____.'  | || | |_________|  | |
| |              | || |              | || |              | || |              | |
| '--------------' || '--------------' || '--------------' || '--------------' |
'----------------'  '----------------'  '----------------'  '----------------'

Version 0.9.44
Copyright(c) 2015 by Thierry Sousbie. All rights reserved.
Author: Thierry Sousbie - tsousbie[a]gmail[dot]com

DICE is a C++ template library designed to solve collisionless fluid dynamics in 6D phase space using massively parallel supercomputers via an hybrid OpenMP / MPI parallelization. This package implements a cosmological and physical VLASOV-POISSON solver for cold systems such as dark matter (CDM) dynamics called ColDICE and based on DICE.

In particular, the following features are implemented in DICE so far:

(I) Installation

The DICE template library itself is located in the include/dice/ directory and features 3 executables:

The Cold dark matter fluid dynamics solver ColDICE is composed of 6 main binaries:

Requirements

Additional requirements for MPI support:

IMPORTANT NOTES:

Installing

Installing is pretty simple, go to the distribution directory ${DIST_PATH} and type:

cd build/
cmake ../ <options>

A report will be displayed with default configuration and indications on how to change options (such as the installation path or path to libraries). Typical options are :

-DCMAKE_PREFIX_PATH=<PATH> indicates default path to all libraries
-DNO_MPI=true to compile without MPI support

Other options are indicated in the report. Problems occuring during configuration that may need fixing are reported as red lines in the report together with a hint on what should be done. If such a problem occurs, you may optionally rerun cmake with different options (it is often advised to clean the cache first):

cd ../; ./cleanup; cd build/
cmake ../ <new options>

Once everything is set, the code is built by running:

make -j N

where N is the number of cores to use in parallel. Upon completion, the resulting binaries are in ${DIST_PATH}/bin. You can also optionally install the package by running:

make install

NOTE:
It may sometimes be necessary to clean-up all the files generated by cmake before changing cmake options (e.g. when a library was found but you want to use one from a different location, or when you want to change the compiler). This can be done by running the cleanup script from the distribution directory

cd ${DIST_PATH}
./cleanup

IMPORTANT NOTES:

Usage

Using DICE in a program is relatively easy as it suffice to include relevant header files in your code. Compiling is also made easy thanks to the dice-config utility which can be used to retrieve the compiler options necessary to the proper functioning of DICE. To compile a program dummy.cpp using DICE, just do:

`${DICE_DIR}/bin/dice-config -all` -I${DICE_DIR}/include dummy.cpp -o dummy

where ${DICE_DIR} is the directory where dice was installed (or the directory where it was compiled if you did not install it). Note that dice-config can also list flags, defines and libraries individually for more complex projects.

(II) Documentation

The documentation is located in the ${DIST_PATH}/doc folder and should contain the following:

(III) Vlasov-Poisson solvers: coldice_xD

Running:

Parameters for coldice programs can be specified directly on the command line preceded by a - marker or directly from a parameter file, in which case the parameter file name must be specified as the first argument, without - marker. They generaly take the form <section>.<parameter> and each parameter can accept one or more arguments, separated by spaces. If a given parameter is not specified, it is given a default value. Information on available parameters, their meaning and default values can be obtained by running:

coldice_nD -help

or by consulting the appropriate section of this manual. Note however that the list of options given by running the code may be more up-to-date, and it might be convenient to save this documentation to a file by running:

coldice_nD -help > parametersDoc.txt

A typical output will take the form of a list of alphabetically ordered parameters:

(...)
Category: 'mesh'
 ->  mesh.allocFactor = 1
   What fraction of the current number of element to reallocate when a pool is full.

 ->  mesh.delta[0] = 2
     mesh.delta[1] = 2
     mesh.delta[2] = 0
     mesh.delta[3] = 0
   Size of the initial bounding box

 ->  mesh.initPartitionTolerance = 1.01
   Tolerance on the load balance of the initial partition
(...)

In this particular example of the 2D version, a parameter -mesh.delta that sets the physical size of the initial bounding box is shown to take four arguments at most with default values 2, 2, 0, and 0. New values (e.g. a size of 1 instead of 2) can be specified on the command line as follows:

coldice_2D -mesh.delta 1 1

where the unspecified remaining two parameter values will take the default value 0.

Note that -help option lists only the options available with default parameters and specifying some parameters may unlock new available parameters. For instance, the option -init.type specifies the initial conditions type and has default value sineWaves with following options:

Category: 'init'
 ->  init.amplitude[0] = 0.2
     init.amplitude[1] = 0.2
   Amplitude of the sine waves

 ->  init.nWaves = 1
   Number of waves along each dimensions

 ->  init.shift[0] = 1
     init.shift[1] = 1
   Sine wave coordinates shift in [0,1]

 ->  init.velFactor[0] = 0
     init.velFactor[1] = 0
   Velocity to displacement ratio

 ->  init.velVector[0] = 0
     init.velVector[1] = 0
   Global velocity vector

Changing -init.Type to cosmo will change these options, which can be unveiled by running coldice_2D -init.Type cosmo -help :

Category: 'init'
 ->  init.fileName = ics_gadget.dat
   Gagdet file name (without .N extension for multiple files).

 ->  init.initTesselationType = ANY
   Topology of the initial grid simplicial tesselation (ALTERNATE,ANY,MINIMAL,REGULAR,USER_DEFINED)

 ->  init.maxMemPerNode = 1
   The maximum amount of memory to use on each MPI node, expressed in GigaBytes.

 ->  init.rowMajor = 1
   Use row major particle ordering if true (i.e. index of particle changes faster with X coordinate). Default is rowMinor.

Note that a summary of the parameter values that will be used for the current run is always given at the end of the output message. A complete list of available parameters is given at the end of this section.

Examples

A few sample parameters defining initial conditions used in the scientific article doc/ColDICE.pdf are given int he ini/ sub directory:

The last example (sineWaves_lowRes.ini) in particular can be ran on a modest workstation as follows:

    coldice_2Dp sineWaves_lowRes.ini -solver.nThreads N -solver.outputDir Dir

With N the number of openMP threads to use and Dir the name of the ouput directory where files will be created (pay attention to the fact that this directory must exist prior to running the code). The program should run for up to a few minutes and produce a bunch of output files that can be visualized with Paraview (http://www.paraview.org).

Note that the phase-space sheet files need to be un-periodized and converted to the vtu format first, which can be done with the following command:

    for i in *.NDnet; do netconv $i -unperiodize -to vtu; done

Output files

There are four types of output files :

Restarting jobs

Restart files are produced at regular intervals specified by option -solver.restartEvery <N>. These restart files can be used to restart a job from were it was stopped with option -restart <file name without .rst extension>. When restarting a job, some parameters may be modified such as the resolution of the grid over which the potential is computed but other are forced to take values given in the initial run such as the mesh resolution. A summay of the "managed" parameters values that will be used for the current run is always given as a header in the output of coldice before starting computations.

Parameters description

An up-to-date list of parameters can be obtained using the -help option, by running coldice_xD -help (see list heareafter). Depending on the selected options, new options may be unlocked, these new options can also be listed by running a similar command line of the form coldice_xD -mycategory.myoption -help. A list of options given by the command coldice_xD -help at the moment this documentation was written follows:

 Category: 'cosmology'
 ->  cosmology.h0 = 1
   Hubble parameter (h0 = H0 / 100 kms-1 Mpc-1)

 ->  cosmology.oB = 0
   Baryonic matter fraction (already included in oM)

 ->  cosmology.oL = 0
   Cosmological constant fraction

 ->  cosmology.oM = 1
   Matter fraction

 ->  cosmology.w = -1
   Equation of state for dark energy

 Category: 'default'
 ->  debug = 0
   Set for debug mode (value 0 to 2)

 ->  debugHook = 0
   Starts an infinite loop on startup to allow hooking a debugger

 ->  debugWait = 0
   Wait for N seconds on startup to allow attaching a debugger

 ->  globalLogDir = 
   The directory where the global log is created

 ->  globalLogFName = globalLog
   The name of the global log file

 ->  ignoreUnusedParams = 0
   Do not exit when encountering unknown params

 ->  initOnly = 0
   Just initialize, report, and exit

 ->  noGlobalLog = 0
   Set to prevent creation of a global log

 ->  restart = 
   The name of a restart file  
   (without the '.rst' or '_${RANK}.rst'(MPI) extension)

 ->  threads_per_node = -1
   The number of openMP threads to use per MPI node (auto set if negative)

 ->  verbose = 3
   Verbosity level (from 0=quiet to 4=debug)

 Category: 'dump'
 ->  dump.amrAt = 
   Sets AMR grid file dumps time interval (fmt: 'start:stop:every').

 ->  dump.amrEvery = 0
   Sets AMR grid file dumps time-steps interval 
   (dump every N timesteps, never if N negative or 0).


 ->  dump.causticsAt = 
   Sets caustics surface sub-mesh file dumps time interval. 
   Format is 'start:stop:every'.

 ->  dump.causticsEvery = 0
   Sets caustics surface sub-mesh file dumps time-steps interval. 
   Dump every N timesteps, never if N negative or 0.

 ->  dump.densityAt = 
   Sets density grid file dumps time interval (fmt: 'start:stop:every').

 ->  dump.densityEvery = 0
   Sets density grid file dumps time-steps interval.
   Dump every N timesteps, never if N negative or 0.

 ->  dump.densityProfileCenter[0] = 0
     dump.densityProfileCenter[1] = 0
   The coodinates of the density profile center. Only used if 
   specifyProfileCenter is set.

 ->  dump.linesAt = 
   Sets lagrangian 1D lines sub-mesh file dumps time interval
   (fmt: 'start:stop:every').

 ->  dump.linesEvery = 0
   Sets lagrangian 1D lines sub-mesh file dumps time-steps interval 
   (dump every N timesteps, never if N negative or 0).

 ->  dump.meshAt = 
   Sets mesh file dumps time interval (fmt: 'start:stop:every').

 ->  dump.meshEvery = 0
   Sets mesh file dumps time-steps interval 
   (dump every N timesteps, never if N negative or 0).

 ->  dump.potentialAt = 
   Sets potential grid file dumps time interval 
   (fmt: 'start:stop:every').

 ->  dump.potentialEvery = 0
   Sets potential grid file dumps time-steps interval 
   (dump every N timesteps, never if N negative or 0).

 ->  dump.radialGridDensityAt = 
   Sets radial density profile from grid file dumps time interval 
   (fmt: 'start:stop:every').

 ->  dump.radialGridDensityEvery = 0
   Sets radial density profile from grid file dumps time-steps interval 
   (dump every N timesteps, never if N negative or 0).

 ->  dump.radialMeshDensityAt = 
   Sets radial density profile from mesh file dumps time interval 
   (fmt: 'start:stop:every').

 ->  dump.radialMeshDensityEvery = 0
   Sets radial density profile from mesh file dumps time-steps interval 
   (dump every N timesteps, never if N negative or 0).

 ->  dump.reset = 0
   Set this parameters to reset any default file dumps 
   (i.e. reset dumps programmed from restart file)

 ->  dump.specifyProfileCenter = 0
   Profile is centered on the center of the bounding box if false, or use user specified 
   coordinates if true (see densityProfileCenter)

 ->  dump.subsetsAt = 
   Sets mesh subsets file dumps time interval (fmt: 'start:stop:every').

 ->  dump.subsetsEvery = 0
   Sets mesh subsets file dumps time-steps interval (dump every N timesteps, never if N negative or 0).

 Category: 'glbDebug'
 ->  glbDebug.meshRefine = 0
   Set to debug mesh refinement.

 Category: 'init'
 ->  init.type = sineWaves
   The type of initial condition: uniformGrid, sineWaves, phasedWave, cosmo, composite
   Each type of initial condition is associated to a different set of parameters in the 'init' section
   listed by running `coldice_xD -init.type TYPE -help`. The following parameters are for 'sineWaves'
   initial conditions type.

 ->  init.amplitude[0] = 2
     init.amplitude[1] = 2
   Amplitude of the sine waves

 ->  init.cosmo = 1
   Whether to use cosmology

 ->  init.delta[0] = 1
     init.delta[1] = 1
     init.delta[2] = 0
     init.delta[3] = 0
   Size of the initial mesh

 ->  init.initTesselationType = ANY
   Topology of the initial grid simplicial tesselation (ALTERNATE,ANY,MINIMAL,REGULAR,USER_DEFINED)

 ->  init.resolution[0] = 64
     init.resolution[1] = 64
   Resolution of the initial mesh (in pixels before tesselation)     

 ->  init.x0[0] = -0.5
     init.x0[1] = -0.5
     init.x0[2] = 0
     init.x0[3] = 0
   Initial coordinates of the mesh lower left corner


 Category: 'mesh'
 ->  mesh.allocFactor = 1
   What fraction of the current number of element to reallocate when a pool is full.

 ->  mesh.delta[0] = 1
     mesh.delta[1] = 1
     mesh.delta[2] = 0
     mesh.delta[3] = 0
   Size of the bounding box

 ->  mesh.initPartitionTolerance = 1.01
   Tolerance on the load balance of the initial partition

 ->  mesh.initPartitionType = KWAY
   The method to use for initial partitionning (KWAY,PH,PH_KWAY)

 ->  mesh.refinePartitionType = ADAPTIVE
   The method to use for repartitionning (ADAPTIVE,KWAY,PH,REFINE_KWAY)

 ->  mesh.repartThreshold = 1.15
   Inbalance factor that triggers repartitionning

 ->  mesh.repartTolerance = 1.01
   Tolerance on the load balance after repartitionning

 ->  mesh.x0[0] = -0.5
     mesh.x0[1] = -0.5
     mesh.x0[2] = 0
     mesh.x0[3] = 0
   Initial coordinates of the bounding box lower left corner

 Category: 'projection'
 ->  projection.accuracyLevel = 0.001
   Required accuracy level for projection (DISABLED). 
   Use D_ENABLE_ACCURACY_CHECKING compile time option to enable.

 ->  projection.anisotropyThreshold = 1e+06
   The simplex anisotropy threshold up to which exact projection is used 
   (expressed as the maximum to minimum simplex extent ratio)

 ->  projection.enableHRModeThreshold = 0
   Fraction of the simplices that are allowed to be reprojected before definitively switching to high 
   resolution projection mode (between 0 and 1, negative number=>always).

 ->  projection.resetHRMode = 0
   Ignore whether high resolution mode was set on a previous run and set it to false. 
   This is only meaningfull when restarting a run.

 ->  projection.useVerticesThreshold = -1
   Minimum refinement level of a simplex above which its vertices are used to project it instead of 
   exact integration. This should be equal to either '-1' or the value of 'solver.maxSimplexLevel'.

 ->  projection.widthThreshold = 1e-07
   The width of an element, expressed as a fraction of the bounding box size, down to which 
   exact projection is used.

 Category: 'solver'
 ->  solver.aEnd = 1
   Final value of the scale factor

 ->  solver.aStart = 0.01
   Initial value of the scale factor

 ->  solver.cflCondition = CFL_RHOMAX
   CFL condition type (CFL,CFL_RHOMAX,NONE,RHOMAX)

 ->  solver.cflGrid = 0.25
   Value of the CFL condition (fraction of a FFT grid pixel size)

 ->  solver.cflRhoMax = 0.01
   Value of the maximum density condition (fraction of an orbit size)

 ->  solver.checkProjectedDensity = 1
   Check the positivity of the projected density (also look for NaN values).

 ->  solver.coarsenEvery = 1
   Force coarsening every N timesteps

 ->  solver.coarsenHysteresis = 0.95
   NOT used

 ->  solver.cosmo = 1
   Use cosmological expansion ?

 ->  solver.dLogA = 0.1
   Maximum value for d(log(a))=da/a.

 ->  solver.dumpInitialMesh = 0
   Dump the mesh right after creating the initial conditions.

 ->  solver.dumpRestartSignalFileName = DUMP_RESTART
   The name of the file to create in the output directory in order to force the creatioàn of a restart 
   file on next time step.

 ->  solver.exportFftWisdom = wisdom.fftw
   The name of a file to which the acquired FFTW wisdom will be exported. Set to 'SKIP' or 'skip' to 
   prevent exportation.

 ->  solver.fastAmrBuild = 1
   If true, refine the AMR grid so that it only fits roughly the resolution of the mesh. This is 
   probably always good enough and much faster ...

 ->  solver.fftGridLevel = 8
   Size of the FFT grid used for the Poisson solver (effective size is 2^fftGridLevel)

 ->  solver.fftWisdom = 1
   Which level of FFTWISDOM to use, in the range {0,1,2,3}={ESTIMATE,MEASURE,PATIENT,EXHAUSTIVE}. 
   The higher the faster, but the slower the initialization ...

 ->  solver.importFftWisdom = wisdom.fftw
   The name of a file from which FFTW wisdom will be imported. Set to 'SKIP' or 'skip' to 
   prevent importation.

 ->  solver.invariantThreshold = 1e-05
   Poincare invariant threshold to trigger refinement

 ->  solver.mass = 1
   Initial total mass

 ->  solver.maxAmrLevel = 8
   Maximum refinement level of the AMR grid (don't touch that, it's already set ;) )

 ->  solver.maxSimplexLevel = -1
   Maximum level of refinement a simplex is allowed to reach.
   (-1  for no maximum, unrefined simplices have level 0)
   See also 'projection.useVerticesThreshold'.

 ->  solver.nThreads = 4
   The number of threads to use per process. This parameter will override the default 'threads_per_node' 
   parameter and will be saved from one run to another.

 ->  solver.noCoarsen = 1
   Set to prevent mesh coarsening

 ->  solver.noRefine = 0
   Set to prevent mesh refinement

 ->  solver.noRepartWeight = 1
   If true, MPI regions weight is directly proportional to the number of simplices.

 ->  solver.noRestart = 0
   Set to prevent from dumping any restart file.

 ->  solver.outputDir = 
   The directory where all files will be put (default is './', directory must exist)

 ->  solver.phSortThreshold = 0.05
   Maximum allowed ratio of randomly distributed to ordered cells before triggering a Peano-Hilbert 
   sort of the local meshes.

 ->  solver.projectionOrder = 1
   Order of the exact mass projection (only 0 and 1 are implemented so far)

 ->  solver.rebuildAmrEvery = 1
   How many timestep to wait before entirely rebuilding the AMR grids. 
   (IGNORED : not implemented yet!)

 ->  solver.refineThreshold = 3
   NOT used

 ->  solver.resimulate = 0
   Set to start a resimulation after the simulation completes

 ->  solver.resimulateEvery = 0
   Not implemented

 ->  solver.restartEvery = 50
   Write a restart file every N timestep (no restart is written if N negative or 0)

 ->  solver.skipInitialPoisson = 1
   Skips the pre-solving of poisson equation for when the CFL conditon is set to CFL_RHOMAX. 
   Instead, measure initial RHOMAX directly on the sheet.

 ->  solver.splitLongestEdge = 0
   Set to split the longest longest edge when refining instead of trying to minimize the invariant

 ->  solver.squeezeMesh = 0
   You don't want to know ;)

 ->  solver.statisticsFileName = statistics.txt
   The name of the file where statistics will be stored

 ->  solver.stopSignalFileName = STOP
   The name of the file to create in the output directory in order to cleanly stop the run.

 ->  solver.symmetry = NONE
   Type of symmetry to enforce.
   (AXIAL0,AXIAL01,AXIAL012,AXIAL02,AXIAL1,AXIAL12,AXIAL2,CENTRAL,CONSTANT0,CONSTANT1,CONSTANT2,NONE,
   PLANAR0,PLANAR01,PLANAR012,PLANAR02,PLANAR1,PLANAR12,PLANAR2)
   THIS OPTION DOES NOT WORK PROPERLY WITH MPI YET !

 ->  solver.timeLimit = -1
   CPU time limit (in hours, will stop before spending it). This is an exact limit (negative for no limit). 
   See also timeLimitSafety

 ->  solver.timeLimitSafety = 0.95
   Fraction of the timeLimit while it is safe to keep computing. Leaves (1-timeLimitSafety)*timeLimit 
   for writing the restart file.

 ->  solver.timingsFileName = timings.txt
   The name of the file where timings will be stored

 ->  solver.volumeThreshold = 1e+10
   Volume threshold to trigger refinement (NOT used)

 Category: 'units'
 ->  units.H = 1
   Hubble parameter (default value in s-1).

 ->  units.length = 1
   Length unit expressed in meters (default: -).

 ->  units.mass = 1
   Mass unit expressed in meters (default: -).

 ->  units.velocity = 1
   Velocity unit expressed in meters (default: -).

Initial conditions

There are currently 5 types of initial conditions implemented in the code. Additionally to these initial conditions, it should be relatively easy to implement your own. If you want to do so, check the IC_dummy.hxx file provided in the ${DICE_DIR}/solvers/ColDICE/init/ directory.

Currently implemented initial conditions are:

Specific parameters are associated to each type of initial conditions. The following list was obtained by running coldice_xD -init.type <TYPE> -help:

TYPE=cosmo: 

 ->  init.fileName = ics_gadget.dat
   Gagdet file name (without .N extension for multiple files).

 ->  init.initTesselationType = ANY
   Topology of the initial grid simplicial tesselation 
   (ALTERNATE,ANY,MINIMAL,REGULAR,USER_DEFINED)

 ->  init.maxMemPerNode = 1
   The maximum amount of memory to use on each MPI node, expressed in GigaBytes.
   Default is 1GB per node.

 ->  init.rowMajor = 1
   Use row major particle ordering if true (i.e. index changes faster with X coordinate). 
   Default is rowMajor.

 ->  init.x0[0] = 0
     init.x0[1] = 0
   Position of the lower left corner of the box

TYPE=sineWaves: 

 ->  init.amplitude[0] = 2
     init.amplitude[1] = 2
   Amplitude of the sine waves

 ->  init.cosmo = 1
   Whether to use cosmology

 ->  init.delta[0] = 1
     init.delta[1] = 1
     init.delta[2] = 0
     init.delta[3] = 0
   Size of the initial mesh

 ->  init.initTesselationType = ANY
   Topology of the initial grid simplicial tesselation 
   (ALTERNATE,ANY,MINIMAL,REGULAR,USER_DEFINED)

 ->  init.resolution[0] = 64
     init.resolution[1] = 64
   Resolution of the initial mesh (in pixels before tesselation)

 ->  init.x0[0] = -0.5
     init.x0[1] = -0.5
     init.x0[2] = 0
     init.x0[3] = 0
   Initial coordinates of the mesh lower left corner

TYPE=phasedWaves

 ->  init.aCross = 0.15
   Time of first shell crossing

 ->  init.delta[0] = 100
     init.delta[1] = 100
     init.delta[2] = 0
     init.delta[3] = 0
   Size of the initial mesh

 ->  init.epsilon = 0.5

 ->  init.initTesselationType = ANY
   Topology of the initial grid simplicial tesselation 
   (ALTERNATE,ANY,MINIMAL,REGULAR,USER_DEFINED)

 ->  init.k = 1
   Perturbation parameter

 ->  init.kp = 1
   Plane wave parameter

 ->  init.resolution[0] = 64
     init.resolution[1] = 64
   Resolution of the initial mesh (in pixels before tesselation)

 ->  init.x0[0] = 0
     init.x0[1] = 0
     init.x0[2] = 0
     init.x0[3] = 0
   Initial coordinates of the mesh lower left corner

TYPE=composite: 

 ->  init.add = none
   The type of component to add (only 'box' is available now) . Can be called several times ...

 ->  init.bBoxDelta[0] = 2
     init.bBoxDelta[1] = 2
     init.bBoxDelta[2] = 0
     init.bBoxDelta[3] = 0
   Size of the bounding box

 ->  init.bBoxX0[0] = -1
     init.bBoxX0[1] = -1
     init.bBoxX0[2] = 0
     init.bBoxX0[3] = 0
   Initial coordinates of the bounding box lower left corner

(IV) Solvers post-reatment from restart files: coldicePost_xD

coldicePost_xD programs are similar to the coldice_xD ones but are adapted to the post-treatment of restart files produced by coldice_xD. They work exactly the same as coldice programs and take the same arguments as coldice_xD plus a few additional ones in the "post" section (i.e. of the form post.<option>). This program is typically run as:

coldicePost_xD -restart <FILENAME without `_XXXX.rst` extension> -post.option...

and all options available in the post section can be listed by running coldicePost_xD -help:

 Category: 'post'
 ->  post.dumpAmr = 0
   Whether to dump AMR files.

 ->  post.dumpDensity = 0
   Whether to dump density grid files

 ->  post.dumpMesh = 0
   Whether to dump the unstructured mesh files

 ->  post.dumpRadialGridDensity = 0
   Whether to dump the radial grid density (from projected grid)

 ->  post.dumpRadialMeshDensity = 0
   Whether to dump the radial mesh density (from the mesh)

 ->  post.dumpSubmesh = 0
   Whether to dump the unstructured mesh subsets files

 ->  post.gridLevel = 8
   Sets the resolution of the density grid

 ->  post.projectionOrder = 1
   Order of the exact mass projection (0 or 1)

 ->  post.radialMeshDensitySamples = 10
   The number of samples to use per overlap for radial mesh density computation

 ->  post.refineMesh = 0
   Refine the mesh using tracers.

(V) Static potential Vlasov-poisson solvers: coldiceStatic_xD

coldiceStatic_xD programs are similar to coldice_xD ones but work with a static potential for which an analytic solution is known. They work exactly the same as coldice programs and take the same arguments plus a few additional ones in the staticSolver section (i.e. of the form staticSolver.<option>). These parameters can also be listed with the "-help" option :

Category: 'staticSolver'
  ->  staticSolver.gravitationalMass = 1
    Total mass generating the static potential

  ->  staticSolver.potential = PLUMMER
    Static potential type (CHAOTIC,PLUMMER,UDF)

and a few parameters are associated to each type of potential, listed in the potential section. For instance, the parameters associated to the CHAOTIC potential are obtained by runnning coldiceStatic_xD -staticSolver.potential CHAOTIC -help:

Category: 'potential'
  ->  potential.Rc = 0.2
    Central radius

  ->  potential.epsilon = 0.5
    Perturbation amplitude (epsilon=0 -> regular orbits)

  ->  potential.q = 0.9
    Excentricity

(VI) Unstructured mesh output files manipulation : netconv

netconv is used to converted unstructured mesh outputs in .NDnet format to other formats and/or unperiodizing the mesh for visualization. A short documentation can be obtained by running netconv -help, and it is typically used to convert files to the vtu format that can be visualized with paraview :

netconv <NDnet filename> [-unperiodize] -to vtu

where the unperiodize option can be specified to explicitly break periodic boundary conditions that would otherwise produce visual artefacts.

(VII) Troubleshooting