If you're tired of reading about me doing NBody simulations, then you might want to skip this post. The interesting thing about this one is not the NBody part, but the LuaJIT part. I've been toying with the idea of learning Lua for awhile now. It doesn't have a lot of the convenient functionalities of Python, but the Just in Time (JIT) compiler promises to be really darn fast. As an added bonus, Lua is used in Torch, a neural net framework that I've played around with in the past. This time however, I wanted to build something in pure Lua, with no help from an external library. I thought it would be instructive to see if I could code up an array object that sort of simulates the behavior of the Numpy array. With this object in hand, it would (presumably) be a piece of cake to code up an NBody simulation. Additionally, I thought it would be fun to do a little speed comparison between my Lua implementation and a Python (numpy) implementation.
