Parallel Computing Toolbox™ 3.3

Product Description

• Parallel Computing Toolbox Key Features
• Programming Task-Parallel Applications
• Programming Data-Parallel Applications
• Working in an Interactive Parallel Environment
• Working in a Batch Environment
• Scaling to a Cluster Using MATLAB® Distributed Computing Server™

Programming Task-Parallel Applications

You can parallelize Monte Carlo simulations and other coarse-grained or embarrassingly parallel problems by organizing them into independent tasks (units of work). Parallel for-loops (parfor-loops) in the toolbox offer one way to distribute tasks across multiple MATLAB workers. Using parfor-loops, you can automatically distribute loop iterations to multiple MATLAB workers, provided the iterations are independent of each other. You can execute parfor-loops interactively from the MATLAB command prompt or execute them offline in a batch environment. The parfor construct manages data and code transfer between your MATLAB session and the workers. It detects the presence of workers automatically and reverts to serial behavior if none are present.

You can also explicitly specify tasks either as MATLAB functions or as MATLAB scripts. You can execute the tasks, when specified as functions, by manipulating task and job objects in the toolbox or, when specified as scripts, by using the batch function.

Using parfor-loop for a task-parallel application. You can use parfor-loops in MATLAB® scripts and functions and execute them both interactively and offline. Click on image to see enlarged view.