I’ve been interested in the Julia language (link) for a while. During the recent holidays I had some plane ride time to write up some simple comparisons with other languages to see how well it fared. I prepared a couple of sets of “benchmarks” in 4 languages: C, Python2.7, Java, and Julia. First I implemented a simple Newton’s method root-finding algorithm for a simple polynomial, running it over the course of 10000 iterations for testing. The second benchmark was an implementation of the BTW Sandpile model that I evolved 10000 steps. Here are the results:

Language comparison chart

Overall, Julia fared pretty well. The code complexity was on par with Python, and the runtimes on par with Java. I think I will continue to mess around with it for a while and see how it does with larger problems.

Julia, as a language, is hit or miss to me. I understand that a lot of the syntax is supposed to be “matlab-y” but I really dislike array indexing starting at 1. I know that’s going to cause me some troubles as I use it more. But, the ability to call C natively is exciting, along with the maturity of the plotting capabilities already.