Installation

Hardware requirements

You need an Nvidia CUDA-capable GPU with compute capability 2.0 or above with the appropriate Nvidia driver. You can check if your GPU is supported on Nvidia’s website.

Software requirements

You need the CUDA Toolkit (version 8.0 or above), which can be installed using Conda (recommended) or from Nvidia. Make sure that the version you install supports your Nvidia driver or upgrade the driver. The driver requirements of each CUDA Toolkit version can be found in the release notes. If you use the CUDA Toolkit not installed by Conda and encounter issues, check that the installation path is configured so that Numba can access it.

Installation

It is strongly recommended to install Disimpy in a virtual environment. The most recent release can be installed with pip:

pip install disimpy

For advanced users

The package can also be installed directly from source (note that you can specify things such as branch and commit):

pip install git+https://github.com/kerkelae/disimpy.git

Automated tests

To confirm that the installed simulator works, you should run tests by executing the following in the Python interpreter:

from disimpy.tests import test_all
test_all()