The security return files are distributed and stored in 100 different directories. Where it gets stored is based upon a simple formula: file index modulus 100, where the file index is the numerical identifier for each security, numbered between 1 to 100,000.
Each data file is in a simple binary format. The upstream process has calculated three source returns for 10,000 future states, as in a 10,000 x 3 matrix. The layout is column-oriented, meaning that the first 10,000 numbers are used for the first return source, the next 10,000 numbers are for the second return source, and so on.
Before we start using distributed computing functions, we must spawn worker processes. Julia comes with a convenient command-line ...
