Release notes

Current development version - v1.4.0 - XXth August 2019

Physics

  • Working implementation of dust growth using Stepinski-Valageas 1997 model (Vericel)
  • updated MCFOST interface for live radiation calculations
  • further improvements to Roche Lobe injection (Worpel)
  • Fixed issue of initial violent response of inner disc - no longer correct orbital velocities for surface density turnover in inner disc
  • default alpha_AV is 1.0 instead of 0.1 when using CONST_AV = yes
  • warning added about particles with zero sound speed
  • preliminary work to incorporate Shen (2012) equation of state for Neutron stars

Setup

  • Bug fixes with disc setup routines
  • Default npart is 10^6 in disc setup
  • better warnings about validity of terminal velocity approximation
  • moved default settings for particle arrays into init_part routine
  • cluster setup reads/writes .setup file

Bugs

  • Issue with zero grain sizes upon restart fixed, now checked for in checksetup
  • Issue with one fluid setups not working on stable branch fixed
  • Numerous bugs fixed with dust growth implementation (Vericel)
  • git version info prints correctly when running test suite
  • now call update_externalforce before checksetup is run to avoid problem with extern_binary
  • Default units changed in galaxies setup to avoid momentum conservation warning

Performance

  • Improved parallelisation of root node construction in kdtree build

Build

  • Nightly code performance (openMP only) now checked automatically

Utils

  • read_array_from_file in utils_dumpfiles can be used to read real*4 arrays not read during read_dumpfile (e.g. luminosity)
  • kernels script updated to Python 3
  • several python scripts (evcat,evcut,evhead,ev2dot) added for messing around with .ev files (#, Liptai)
  • phantom2hdf5 added to convert dump files to hdf5 format (Mentiplay, Liptai)
  • moddump to remove particles inside/outside some radius (Vericel)
  • disc analysis utility now assumes that the disc is around the first sink if sinks are present
  • combinedustdumps utility to stack different grain sizes from single-grain calculations now works with automatic memory allocation

Other

  • less verbose output during memory allocation
  • update_test_scores routine used to avoid repeated code in test suite
  • optional HDF5 output for easy reading of dump files in Python via Plonk (Mentiplay, Liptai)
  • automatic correction of “if(” to “if (” by format-bot

v1.3.0 - 22 Feb 2019 - 4d45cb3

Physics

  • Multigrain dust simulations with multiple large grains now possible (Mentiplay). This complements the multigrain method used for small grains, but simulating small and large grain populations simultaneously is not yet fully functional
  • Further updates to dust growth algorithms (Vericel)
  • Much improved wind injection routines (Price, Siess)
  • Improvements to Roche lobe injection module (Worpel)
  • Injection modules can now provide an additional timestep constraint where needed
  • One fluid dust uses method of Ballabio et al. (2018) to prevent negative dust fractions
  • can now set a maximum density after which the simulation will end, also dtmax will dynamically decrease/increase if density increases too rapidly (Wurster)
  • removed obsolete and unused etamhd fixed resistivity variable
  • reduced timestep from physical viscosity force by factor 0.4: this has been found to lead to much better convergence of disc simulations that use this method (Nixon)

Bugs

  • bug fix with momentum conservation in two fluid dust-gas drag when ISOTHERMAL=yes
  • array bounds error in analysis_tde fixed
  • bugfix in read options for externbinary module

Tests

  • test for momentum and energy conservation in two fluid dust-gas drag
  • code performance is now checked nightly against a suite of benchmarks
  • sends error code to system if a fatal error happens (Pinte)
  • added check on the conservation of angular momentum with dust/gas

Setup

  • Binary disc setup uses Farris et al. (2014) locally isothermal equation of state for discs around more than one star
  • Disc setup routine modularised and made more general (Mentiplay)
  • gwdisc setup now allows disc inclination (Pereira et al. 2019)
  • setup_star given fairly major restructure so logic is clearer; more cleanly split interactive from non-interactive parts
  • Flyby setup updated with the following roll angle convention: incl=0 => prograde orbit (disc and perturber anti-clockwise; incl=180 => retrograde orbit (disc anti- and perturber clockwise). See Cuello et al. 2019
  • minor fixes to dustyshock and dustywave setups (Hutchison)
  • binary_w in setup_disc is now 270 degrees by default
  • asteroidwind setup added
  • added option to setup a settled dusty disc, working with both one and 2 fluid (Dipierro)

Build

  • version number and git sha now written to dump file headers
  • memory is now allocated at runtime for main arrays in Phantom (Chan). This avoids the need to recompile with MAXP= when you change the particle number. Only applies to main phantom binary at present, not to phantomsetup.
  • many compiler warnings fixed
  • cleanup of evolve module
  • obsolete preprocessor flags -DSORT_RADIUS_INI T and -DDUSTFRAC deleted
  • you can now supply JOBNAME= when making job scripts with make qscript, otherwise it continues to choose delightful random words

Analysis

  • Multigrain post-processing works properly with MCFOST
  • phantomevcompare will not duplicate data when merging files
  • further integration with MCFOST
  • analysis disc planet prints the effective tilt between the inner and outer disc (Nealon)
  • disc analysis now defaults to sorting particles by cylindrical radius - this should fix any discrepancies that may have been occurring. Deliberately made it very hard not to chose this option (Nealon)
  • disc analysis now returns the total angular momentum components as well (Nealon)
  • precession files: these can now be made even if the first file input is not the first file of the simulation (Nealon)
  • utils_disc now handles an eccentric disc - bins are defined by semi-major axis, not by radius (Nealon)
  • analysis_dustydis c
  • Added check Ltot!=0 to prevent NaNs in the output (Ragusa)
  • moddump_extenddis c implemented to extend an existing disc simulation in radius (Nealon)
  • disc scale height now calculated from particle positions but works perfectly with a warped disc (Nealon)

Other

  • phantom outputs helpful error message if .setup file is given on command line instead of .in file

v1.2.0 - 20 Jun 2018 - d339b10

This release corresponds to the accepted version of the Phantom paper (v2 on arXiv). Changes compared to v1.1.0:

Physics

Build

  • SYSTEM=ozstar added

v1.1.0 - 5 Apr 2018

Physics

  • Helmholtz equation of state implemented (Tricco)
  • preliminary work on dust growth (Vericel)

Bugs

  • bug fix with magnetic fields on boundary particles
  • bug fix with incorrect fatal error on centre of mass non-conservation
  • angular momentum now conserved during sink particle accretion (#17, Wurster)
  • issues with git-lfs fixed
  • bug fix with write of B-field to small dump files

Tests

  • setupbot: Nightly checks that phantomsetup does not require unspecified user input

Setup

  • better defaults in several setups so we pass setupbot checks
  • set_slab utility routine added for 2D-in-3D setups

Build

  • SYSTEM=raijin added

v1.0 - 13 Mar 2018

Physics

  • working MPI implementation (Chan)
  • more robust algorithm for one fluid dust (Ballabio+ 2018)
  • dust algorithm (one fluid/two fluid) chosen at runtime not compile time
  • particle waking with individual timesteps re-implemented (Wurster; 45fae9b)
  • universal disc setup routine (Mentiplay)
  • setup added for flyby simulations (Mentiplay, Cuello)
  • CO cooling implemented (Glover)
  • magnetic field evolves B/rho rather than B (Tricco, Price)
  • stellar wind routine works out-of-the-box (Toupin)
  • improvemements to Galactic Centre winds and cooling (Russell, Price)
  • NICIL updated to v1.2.3 (Wurster)

Bugs

  • bug with drag in two fluid dust-gas when hj > hi fixed (Dipierro)
  • updates/bug fixes to MESA Equation of state tabulation
  • bug fix with energy conservation with softened sink particles
  • bug fix with self-gravity + multiple particle types

Tests

  • nightly checks for non-ideal MHD added
  • self gravity checked for all particle types
  • testsuite checked nightly with MPI

v0.9 - 14 Feb 2017

This is the first public release of Phantom, alongside arXiv paper.

Contains:

  • hydro
  • sink particles
  • self-gravity
  • MHD
  • dust (two fluid and one fluid)
  • ISM chemistry and cooling
  • physical viscosity
  • non-ideal MHD
  • external forces including corotating frame, Lense-Thirring precession, P-R drag, fixed binary