Tips for installing NEST with conda¶
This page provides a series of recommendations for installing pre-built NEST with conda or to set up conda environments for building NEST and NEST documentation.
Basic conda setup¶
Choice of conda base installation¶
We test NEST in conda environments using Miniconda installations and thus recommend that you do the same. The recommendations that we provide here will also likely work with a full-sized Anaconda installation, but we can only provide limited support for this.
You can either install
Miniconda from https://docs.conda.io/en/latest/miniconda.html
Miniforge from https://github.com/conda-forge/miniforge
For Apple systems with an M1 chip, you must at present use Miniforge and
arm64 (Apple Silicon) installer to create a conda environment
that will support native builds of NEST.
Keep your base environment empty¶
Your base environment should be as empty as possible in order to avoid conflicts with other environments. Always install packages only in the new environments (don’t worry about duplicates, conda will link the packages if they are used in multiple environments, and not produce disk eating copies).
Installing NEST with Conda¶
We provide pre-built versions of NEST on Conda Forge <https://anaconda.org/conda-forge/nest-simulator/files>. Follow these instructions to install NEST from Conda Forge.
Creating a Conda environment for running and building NEST¶
If you want to compile NEST yourself, you can create an environment containing all necessary software for running and building NEST by executing the following command from the NEST source directory
conda env create -f extras/conda-nest-simulator-dev.yml
This will create an environment called
nest-simulator. If you would like to give the environment
a different name, use
conda env create -f extras/conda-nest-simulator-dev.yml -n MYNAME
Get a good overview¶
Obtain a good overview of which packages are installed where. You can use
conda env export -n base and
conda env export -n yournestenv
yournestenv name with whatever you chose). Make
sure each environment contains all dependencies. One way to make
this obvious would be to reduce conda stack to
0 (conda documentation
and/or to a certain degree by not auto-activating the base environment (conda documentation
Then packages from base do not ‘leak’ into your new environments.
Packages from your system will usually also be available in your conda environment and may cause similar conflicts.