Harrison Roberts, MathWorks
Parallel Computing Toobox™ lets you solve computationally and data-intensive problems using multicore processors, GPUs, clusters, and clouds. Parallel computing is ideal for problems such as parameter sweeps, optimizations, and Monte Carlo simulations. High-level constructs in Parallel Computing Toolbox, such as parallel for-loops and special array types, let you parallelize MATLAB® applications without CUDA or MPI programming. You can also use the toolbox to run multiple Simulink® simulations of a model in parallel. Without changing the code, you can run the same applications on a cluster or cloud using MATLAB Parallel Server™.