AMUSE¶

AMUSE (Astrophysical MUltipurpose Software Environment) originates some ideas from its predecessors: ACS, StarLab and NEMO, but uses the python language. Another feature of AMUSE is that python is also the glue shell between legacy codes that can orchestrate simulations taking components from different codes, whereas in NEMO legacy codes have a NEMO CLI interface, at best.

For seasoned AMUSE users, here we highlight some differences between the two, and give some examples how to achieve the same task in NEMO and AMUSE.

Differences¶

Shell: NEMO uses a Unix shell, AMUSE uses python (ipython, jupyter, …), although commandline scripts do exist. A neat way to start an interative amuse is via: ipython –profile amuse
Community Code: Both packages maintain a tight connection to legacy software and community codes. You can find them in $AMUSE/src/amuse/community and $NEMO/usr resp., though the latter has some supporting script in $NEMO/src/scripts/mknemo.d
Units: NEMO uses dimensionless values, and units are implied. Most programs actually use virial units (a.k.a N-body units, or Henon units) where G=1, but there are a few programs (e.g. galaxy, nbodyX) that use other units. The units(1NEMO) tries to help you converting. AMUSE (optionally?) attaches units to numbers , using a python trick, e.g.

from amuse.units import units

mass   = 1.0 | units.MSun

astropy users might be a bit baffled, since this looks very different. But

m1 = mass.as_astropy_quantity()

will look more familiar. In pure astropy it might look as follows:

from astropy import units as u

m = 1.0 * u.solMass
m2 = m.to(u.kg).value

Examples: Creating a Plummer sphere¶

Here we create a Plummer sphere, in virial units, in NEMO, and display an X-VX projection on the sky in a shell session:

source /opt/nemo/nemo_start.sh

mkplummer p100 100
snapplot p100 xvar=x yvar=vx

or in the style of using pipes this can be a one liner. Here is that example, and a few followups with grey scale and contour plots:

mkplummer - 100 | snapplot - xvar=x yvar=vx
mkplummer - 10000 | snapgrid - - xvar=x yvar=vx   | ccdplot  -
mkplummer - 10000 | snapgrid - - xvar=x yvar=vx   | ccdplot  - 0.01,0.1,0.3,0.6,0.9
mkplummer - 10000 | snapgrid - - xvar=x yvar=vx   | ccdsmooth -  - | ccdplot  - 0.01,0.1,0.3,0.6,0.9

And in AMUSE the following python session can do something similar:

Todo

switch to using matplotlib

from amuse.units import units
from amuse.units import nbody_system
from amuse.ic.plummer import new_plummer_sphere

convert_nbody = nbody_system.nbody_to_si(100.0 | units.MSun, 1 | units.parsec)
plummer = new_plummer_sphere(1000, convert_nbody)

# gnuplot
plotter = Gnuplot.Gnuplot()
plotter.splot(plummer.position.value_in(units.parsec))

# matplotlib
x=p[:,0].value_in(units.pc)
y=p[:,1].value_in(units.pc)
plt.plot(x,y,'ok')

The AMUSE manual has some NEMO I/O examples.

Installation¶

AMUSE can usually be installed easily as follows (as a user):

pip install amuse

but this can take a while as it finds the right dependencies and needs to compile massive amounts of code. Some of these can easily fail if you don’t have the correct prerequisites (e.g. MPI).

A potentially faster way is to first install the AMUSE framework and then the selected module(s):

pip install amuse-framework
pip install amuse-bhtree amuse-hermite amuse-seba amuse-brutus

There are many more details in the AMUSE installation manual.

As a developer, it is easier to install the source code via github, and follow the following recipe somewhere in your workflow (skipping details where your python/virtual environment is)

git clone https://github.com/amusecode/amuse
cd amuse
./setup develop amuse-framework
./setup develop bhtree
./setup test    bhtree

this is the recommended way for NEMO users.

The directory $NEMO/usr/amuse has some example supporting code. A quick commandline script to generate a Plummer sphere and convert it back to NEMO is the following:

amuse_convert.py -o plummer.txt -n 1024
tabcols plummer.txt | tabtos - plummer.snap block1=mass,skip,vx,vy,vz,x,y,z nbody=1024