Installation
Dependencies
dadi depends on a number of Python libraries. The absolute dependencies are
- Python 3
- NumPy
- SciPy
- nlopt
It is also recommended that you install
- matplotlib
- IPython
The easiest way to obtain all these dependencies to install the Anaconda Python Distribution. You'll need to separately install nlopt from conda-forge. And the easiest way to install dadi is then via conda
, using the command conda install -c conda-forge dadi
. dadi can also be installed via pip
, using the command python3 -m pip install dadi
.
GPU computing
dadi can be sped up substantially by running on a CUDA-enabled Nvidia GPU.
To enable this functionality, you will need to install the CUDA Toolkit.
After install the CUDA Toolkit, you will then need to install PyCUDA and scikit-cuda.
Both of these can be installed from the Python Package Index using pip, python3 -m pip install pycuda
and python3 -m pip install scikit-cuda
.
Installing from source
dadi can be easily installed from source code, as long as you have an appropriate C compiler installed. (On OS X, you'll need to install the Developer Tools to get gcc. On Windows, you'll need to install the Microsoft Visual Studio to get C/C++ builder.) To do so, first unpack the source code tarball, unzip dadi-<version>.zip
In the dadi-<version>
directory, run python setup.py install
. This will compile the C modules dadi uses and install those plus all dadi Python files in your Python installation's site-packages
directory. A (growing) series of tests can be run in the tests
directory, via python run_tests.py
.
If you want the latest and greatest from dadi, clone the Bitbucket repository. You can then create a local install using python setup.py develop
. This will ensure that when you edit pull revisions or edit the source code, changes are reflected immediately, without requiring a separate install step.