Quantcast

Parallel Computing Toolbox 

Tutorials

This set of tutorials introduces MATLAB users to MathWorks parallel computing tools. Through code examples, users learn to run parallel MATLAB applications on multicore desktops or clusters.

Completion time: 1-2 hours
Required Products: Parallel Computing Toolbox
MATLAB Release: MATLAB 7.10 (R2010a)

Product Landscape

Parallel Computing Self Tutorial: Product Landscape 2:45 
Use parallel computing tools on a multicore computer or a cluster of computers.

In this video you will learn about:

  • Products required for parallel computing with MATLAB
  • Problem types suitable for solving with parallel computing
  • MATLAB and Simulink products with parallel functions

Prerequisites

Parallel Computing Self Tutorial: Prerequisites 2:01 
Review hardware and product requirements for running the parallel programs demonstrated in this tutorial.

In this video you will learn about:

  • Hardware requirements for parallel computing
  • Requirements for this tutorial series
  • How to test your desktop computer for readiness

Quick Success Example

Parallel Computing Self Tutorial: Quick Success Example 6:13 
Speed up a simple program using a parallel for-loop.

In this video you will:

  • See short examples for using parfor to accelerate computations
  • Set up a problem to run in batch or offline mode

Parallel Computing Concepts

Parallel Computing Self Tutorial: Parallel Computing Concepts 5:19 
Review the concepts required to write and run parallel MATLAB programs.

In this video you will learn the terminology used in these tutorial videos.

Parfor or Jobs and Tasks

Parallel Computing Self Tutorial: parfor or Jobs and Tasks 3:20 
Select the programming construct that is best suited for running your program in parallel.

In this video you will learn about the advantages and limitations of using parfor, as well as using jobs and tasks for distributing computations.

Using parfor Loops

Parallel Computing Self Tutorial: Using parfor Loops 6:16 
Use parfor loops to run programs in parallel. Apply techniques to obtain maximum speedup.

In this video you will learn about:

  • How parfor classifies variables and transmits them to workers
  • Useful techniques to work around some of parfor restrictions
  • Factors governing the speedup of parfor loops

Working with Jobs and Tasks

Parallel Computing Self Tutorial: Working with Jobs and Tasks 10:30 
Run programs in parallel using jobs and tasks. Obtain speedup and avoid common programming pitfalls.

In this video you will learn how to use job and task objects to launch computations on worker processes.

Working with Schedulers

Parallel Computing Self Tutorial: Working with Schedulers 5:12 
Increase speedup further using schedulers to run parallel programs on clusters of computers.

In this video you will learn about:

  • Interacting with schedulers (local, MATLAB job scheduler, and third party schedulers) to launch computations
  • Ensuring your code and data files are available to workers