There are many MATLAB sites on the web. Here are some starting points.

• The MathWorks is the first site you should visit. This is the home of MATLAB.
• The McCormick MATLAB Support site, for NU Engineering students. See their Introduction to MATLAB for Engineering Students (PDF), even if you are not.
• Kellogg Research Computing MATLAB page for NU Management students.
• University of Cambridge Department of Engineering MATLAB page. Good advice, good links.
• Tutorials of an introductory nature.
• The Econometrics Toolbox: by James P. LeSage.
• Links to other MATLAB things by Spatial Econometrics.
• MIT Lecture Notes 10.001, "Matlab, Introduction" (PDF). Forty-five slides introducing you to MATLAB.

Getting Started with MATLAB (PDF) is a comprehensive introduction to MATLAB, its desktop, computational and graphical tools.

MATLAB Programming Tips (PDF) is a concise collection of explanations and advice to help you program MATLAB.

Techniques for Debugging MATLAB M-files is a must-read for any MATLAB user.

The MathWorks List of Available Technical Notes for MATLAB.

You can greatly improve the performance of your MATLAB codes by always vectorizing your code.

• MATLAB Vectorisation Tricks, from the University of Cambridge Department of Engineering.
• The MathWorksTech Note 1109, Code Vectorization Guide.
• The MathWorks Improving Performance and Memory Usage.
• The MathWorksTech Note 1605, MEX-files Guide. You can call C or FORTRAN routines directly from MATLAB.

Vectorized code makes it easy to exploit the parallel compute engine in an NVIDIA CUDA-enabled GPU. First, make sure that you have a CUDA-enabled GPU installed in your PC. If not, you might be able to purchase an NVIDIA GPU and install it yourself. Then install and test NVIDIA CUDA, the software you need to work with the GPU. With the preliminaries completed, it's easy to use MATLAB to program your GPU, starting with vectorized code. GPULib (free for academic use) provides bindings that help you use standard MATLAB code to access the GPU. The Jacket Engine for MATLAB also enables standard MATLAB code to run on any CUDA-capable NVIDIA GPU.

Compile MATLAB Programs. With this technique, you can compile and run MATLAB M-files as multiple concurrent standalone applications (which do not require license server access).

John D'Errico's Optimization Tips and Tricks, for the Optimization Toolbox, Linear and Nonlinear Regression is oriented toward both old and new users.

Additional MATLAB Toolboxes

Other sources of toolboxes include:

• The MATLAB Central File Exchange
• Mathtools.net
• MathPros
• The Mastering MATLAB site
• StatLib - Software and Extensions for MATLAB

Northwestern University    Academic Technologies    bef@northwestern.edu    Last updated July 27, 2009

Services | Get Connected | Support | Educational Resources | NUIT
Students | Staff/Faculty | Visitors/Media | Site Map | Online Directory
Northwestern Home | Northwestern Calendar: Plan-It Purple
Northwestern Sites A-Z | Northwestern Search
Information Technology  Rebecca Crown Center  633 Clark Street  Evanston, Illinois 60208 
E-mail: it-feedback@northwestern.edu
Last updated   World Wide Web Disclaimer and University Policy Statements 
© 2004 Northwestern University