Quickstart

Installation

The installation via PyPi is the simplest method

pip install MCEq

Optionally, one can (and is encouraged to) accelerate calculations with BLAS libraries. For Intel CPUs the Math Kernel Library (MKL) provides quite some speed-up compared to plain numpy. MCEq will auto-detect MKL if it is already installed. To enable MKL by default use

pip install MCEq[MKL]

More speed-up can be achieved by using the cuSPARSE library from nVidia’s CUDA toolkit. This requires the cupy library. If cupy is detected, MCEq will try to use cuSPARSE as solver. To install MCEq with CUDA 10.1 support

pip install MCEq[CUDA]

Alternatively, install cupy by yourself (see cupy homepage).

Supported architectures:

• Linux 32- and 64-bit (x86_64 and AArch64)

• Mac OS X

• Windows

Note

pip installations of scipy on Windows may be faulty. If scipy throws errors on import, use Miniconda

Upgrading

For installations with pip, upgrading the code and data tables can be done with

pip install MCEq --upgrade

In case of major updates the database file will be updated on first import and the old one will be removed. For installations from source, pull the latest release or master branch. The database file will be updated automatically as well.

Building MCEq from source

To modify the code and contribute, the code needs to be installed from the github source

git clone https://github.com/afedynitch/MCEq.git
cd MCEq
pip install -e .

This will build and install MCEq in editable mode, and changes to this source directory will be immediately reflected in any code that imports MCEq in the current python enviroment.

Running MCEq

Open an new python file or jupyter notebook/lab

from MCEq.core import config, MCEqRun
import crflux.models as crf
# matplotlib used plotting. Not required to run the code.
import matplotlib.pyplot as plt


# Initalize MCEq by creating the user interface object MCEqRun
mceq = MCEqRun(

    # High-energy hadronic interaction model
    interaction_model='SIBYLL23C',

    # cosmic ray flux at the top of the atmosphere
    primary_model = (crf.HillasGaisser2012, 'H3a'),

    # zenith angle
    theta_deg = 0.
)

# Solve the equation system
mceq.solve()

# Obtain the result
# Multiply fluxes be E**mag to resolve the features of the steep spectrum
mag = 3
muon_flux = (mceq.get_solution('mu+', mag) +
             mceq.get_solution('mu-', mag))
numu_flux = (mceq.get_solution('numu', mag) +
             mceq.get_solution('antinumu', mag))
nue_flux = (mceq.get_solution('nue', mag) +
            mceq.get_solution('antinue', mag))

# The lines below are for plotting with matplotlib
plt.loglog(mceq.e_grid, muon_flux, label='muons')
plt.loglog(mceq.e_grid, numu_flux, label='muon neutrinos')
plt.loglog(mceq.e_grid, nue_flux, label='electron neutrinos')

plt.xlim(1., 1e9)
plt.xlabel('Kinetic energy (GeV)')
plt.ylim(1e-6, 1.)
plt.ylabel(r'$(E/\text{GeV})^3\,\Phi$ (GeV cm$^{-2}$\,$s$^{-1}\,$sr$^{-1}$) (GeV)')
plt.legend()
plt.show()

Warning

Due to a bug found in the decay table generation for the “new” MCEq versions there are some larger changes for the lowest energies at tens of GeV. To estimate if your computations may be affected check out v1.2.0 vs. v1.1.3.