brainhuck

A Haskell implementation of a Brainfuck interpreter.

The interpeter is implemented as an executable (installed with stack install in the directory), which can take the path of Brainfuck source code, or using the --stdin flag, execute code given as an argument to the interpeter.

Usage examples:

brainhuck path/to/program.b
brainhuck --stdin ">>>>+"
brainhuck program.b -s 100 > debug.txt

Flags/Options:

-s --size # memory size (default: 500 cells)
--stdin # execute code given as an argument

I have yet to to document the source code at least to some extent, nevertheless I think it's fairly readable, though maybe not so idiomatic, since I'm still a Haskell beginner-intermediate.

There are two modules for the interpeter in this project. Their only difference is that one uses Boxed Vectors for the memory, and the other uses Unboxed Vectors. Its important to note, that there are two versions of the Unboxed version. One akin to the one with Boxed Vectors, and another that enables optimizations, by means of simplifying the actions (i.e. for "+++++", instead of performing one at a time, every adyacent '+' fuses into a single action). The difference in performance can be seen in the bench.html generated by criterion. Depending on the program, the optimizations will have varying impact on the final performance.

Image of the benchmarking results

It is important to note that the brainhuck executable uses the optimized version.

This interpreter allows for different datatypes to be used as the memory used, only requiring for each datatype a couple of typeclasses, and few more code.