A Practical Guide to Parallel Computing in Macroeconomics
This repository contains the source code referenced in the paper A Practical Guide to Parallel Computing in Macroeconomics by Jesús Fernández-Villaverde and David Zarruk Valencia.
Abstract from the paper
Parallel computing opens the door to solving and estimating richer models in Economics. From dynamic optimization problems with high dimensionality to structural estimation with complex data, readily-available and economical parallel computing allows researchers to tackle problems in Economics that were beyond the realm of possibility just a decade ago. This paper describes the basics of parallel computing for economists, reviews widely-used implementation routines in
CUDAand compares performance gains using as a test bed a standard life-cycle problem such as those used in macro, labor, and other fields.
Makefile contains the compilation flags used in Linux and can be used to execute the codes in every language.
Cpp_main.cpp: C++ code for OpenMP
CUDA_main.cu: CUDA code
Julia_main_parallel.jl: Julia code
Julia_main_pmap.jl: Julia code
Julia_threads.jl: Julia code with @threads parallelization
Matlab_main.m: Matlab code
MPI_host_file: MPI host file
MPI_main.cpp: C++ code for MPI
Python_main.py: Python code
Python_numba_main.py: Python code with numba parallelization
Rcpp_main.cpp: C++ code for Rcpp package in R
Rcpp_main.R: Rcpp code
R_main.R: R code
Makefile: Makefile to execute codes