Windows binary installers are available here for garlicsim and here for garlicsim_wx.

For a source installation: Download the repo, then run setup.py install twice, once in the garlicsim folder and once in the garlicsim_wx folder.

To start the GUI:

import garlicsim_wx
if __name__ == '__main__': # Needed for multiprocessing
    garlicsim_wx.start()

Core requirements:

• Python, version 2.5 or 2.6. If you're new to Python, download version 2.6. If you use Python 3.1 or above, use this fork of GarlicSim instead.
• Either setuptools or Distribute.

Some bundled simpacks require:

• NumPy and SciPy.

Recommended, but not mandatory:

• Psyco.
• on Windows only: Python for Windows Extensions.
• Backport of the multiprocessing module. Relevant only for users of Python 2.5.

Gui requirements:

• wxPython (Not needed for non-gui usage.)

If you wish, it's possible to just run the gui and play with it without installing anything. To do so, download the repo and run the run_gui.py file in the root folder.

What to do in the GUI? Select File -> New. Choose one of the simulation packages, press Ok. A dialog will pop up, press Ok. Double click the seek bar to toggle playing.

GarlicSim is an ambitious open-source project in the field of scientific computing, specifically computer simulations. It attempts to redefine the way that people think about computer simulations, making a new standard for how simulations are created and used.

GarlicSim is a platform for writing, running and analyzing simulations. It is general enough to handle any kind of simulation: Physics, game theory, epidemic spread, electronics, etc.

When you're writing a simulation, about 90% of the code you write is boilerplate; Code that isn't directly related to the phenomenon you're simulating, but is necessary for your simulation to work. The aim of GarlicSim is to write that 90% of the code once and for all, and to do it well, so you could concentrate on the important 10%.

GarlicSim defines a new format for simulations. It's called a simulation package, and often abbreviated as simpack. For example, say you are interested in simulating the interaction of hurricane storms. It is up to you to write a simpack for this type of simulation. The simpack is simply a Python package which defines a few special functions according to the GarlicSim simpack API, the most important function being the step function.

The beauty is that since so many simulation types can fit into this mold of a simpack, the tools that GarlicSim provides can be used across all of these different domains. Once you plug your own simpack into GarlicSim, you're ready to roll. All the tools that GarlicSim provides will work with your simulation.

Additionally, GarlicSim will eventually be shipped with a standard library of simpacks for common simulations, that the user may find useful to use as-is, or with his own modifications.

For a more thorough introduction to how GarlicSim works, check out the ** Introduction to GarlicSim ** - Though not yet complete, it goes deep into the principles of GarlicSim and how to work with it.

GarlicSim itself is written in pure Python. The speed of simulations is mostly dependent on the simpack's performance - So it is possible to use C code in a simpack to make things faster.

Current screenshot, showing the Game of Life simulation package shipped with the program:

Mockup:

The main mailing list is garlicsim@librelist.org.

The development mailing list is garlicsim.dev@librelist.org.

To subscribe just send an email. These lists are hosted by librelist, which is currently slightly experimental.

This repository contains two packages, garlicsim, which is the core logic, and garlicsim_wx, which is the wxPython-based GUI.

The garlicsim package is the important one, and its code is well-organized and very readable. It is distributed under the LGPL2.1 license.

garlicsim_wx is in a far less mature state than garlicsim. Also, it is not licensed as open source. (Though the source code is available and not obfuscated.) I have not yet decided if the gui will be developed as an open source project or as commercial software, so in the meantime it is officially closed source.

Both packages are copyright 2009 Ram Rachum.

GarlicSim supports Python versions 2.5 and up, not including Python 3.x.

There is a separate fork of GarlicSim that supports Python 3.1. Take note though that it does not contain a GUI, because wxPython does not support Python 3.1.

Garlicsim is at version 0.2.x, which is an alpha release. It is still very experimental, and there are probably many bugs. If you run into any trouble, let us know immediately in the mailing list.

At this experimental stage of the project, backward compatibility will not be maintained. However, I will be available to assist in issues related to backward compatibility.

People often ask this; probably because they do not fully believe it when they read GarlicSim's description saying that it can handle any kind of simulation. Well, it can. It is very general.

Then people ask, if it is so general, how is it useful? There are two answers to that:

1. It is very useful. There are many many features that are common to all kinds of simulations, and which you really don't want to spend time writing. A few examples of such features: Saving the simulation to file. Browsing the timeline of a simulation like a movie clip while it is still crunching in the background. The option to make manual changes to the simulation world, observe their effects on how the simulation unfolds, and then to be able to revert to the timeline in which the changes were not applied. Changing the arguments to the step function on the fly, etc.

2. If you are interested in only a specific subset of simulations -- say, simulations of solid bodies in Physics -- Then it will be the wisest to write a framework for that within the framework of GarlicSim. Indeed, part of the work on GarlicSim will include writing these kind of sub-frameworks for the common categories of simulations (e.g., a framework for physics, a framework for game theory, etcetera.)

No. GarlicSim doesn't contain any mathematical tools, or any algorithms to be used in simulations. These things are related to the content of the simulation, which is none of GarlicSim's business. GarlicSim handles the organization, or "bureaucracy" of the simulation, so you can concentrate on the content. If you need mathematical tools for your simulation there are many Python projects that provide them, and you may use them in your simpack that you run with GarlicSim.