bbengfort / fastlife

Experimenting with C code and Python bindings

Fast Life

Fast Life is an experiment in Python simulation performance. Python has many excellent simulation frameworks including SimPy and MESA, which make it easy to conduct simulations for research and experimental purposes. Fast Life is not intended to be a simulation framework in the same way, instead Fast Life is designed to answer the question: "Can Python be used to create extremely large scale simulations"? To this end, Fast Life implements a seemingly simple simulation: Conway's Game of Life -- but scales it up to massive proportions. We then explore several different implementations including:

1. Pure Python Sequential
2. Pure Python Multiprocessing
3. C-bindings Sequential
4. C-bindings Parallel

Each of these implementations will be benchmarked with significantly sized worlds and the same random seeds to determine how their performance scales. Please see fastlife.readthedocs.io for more details and the complete write up on the documentation.

Quick Start

Install the package and the command line tool using the Python package manager as follows:

$ pip install fastlife

Alternatively, if you're interested in development, you can download the repository, cd into it and install locally with:

$ git clone https://github.com/bbengfort/fastlife.git
$ cd fastlife
$ pip install -e .

Note that if you're developing, you should probably set up a virtualenv and all of that good stuff before doing this step. Once done, you should have a CLI script fastlife installed on your path, check it out as follows:

$ fastlife --help

To run a simulation and see the animated action, you would use fastlife run -a - you can play with the various settings and commands to see different implementations of the simulation.