In the spring of 2008 we invested in the MATLAB Parallel Computing Toolbox (PCT), which was called the Distributed Computing Toolbox (DCT) at the time. One of the main attractions of this product was the ease of implementation - the ability to "parallelize" your code with minimal changes. My applications thus far have all been of the "distributed" (i.e. embarassingly parallel) kind. Generally speaking, I would agree that the MATLAB toolbox makes it very convenient to take advantage of multiple processors with this type of problem. The details, however, are never quite as simple as you would hope. On this page I am listing some examples of codes that work, and some examples of problems I have encoutered with the PCT.

If you have comments, corrections etc.. on these examples you can contact me at "art.gleason(at)miami.edu".

• My experience with creating figures within parfor has been that it works, but I can not control the resolution of figures created within parfor (May 2008). (Update Jun 2008)
• Problem with user-defined objects in parfor (May 2008). Note: this appears to be resolved with R2008a. (June 2008)
• Problem with sliced variables and function calls (Mar 2008)
• A trivial example of parfor to get started (Jan 2008)

Return to Art Gleason Home

Last Updated: June 2008