Natick, MA – A new version of its Parallel Computing Toolbox is now available from Mathworks. Version 4.2 provides an improved distributed array construct that lets MATLAB users directly access from a MATLAB session data that is stored on multicore computers or computer clusters.
In addition, key algorithms in Statistics Toolbox and Communications Toolbox now execute faster when run in conjunction with Parallel Computing Toolbox.
With the new capabilities in Parallel Computing Toolbox, distributed SOF arrays and the improved set of MathWorks parallel computing tools that work with them let MATLAB users easily manipulate large data sets that reside on a computer cluster or multicore computer without significant changes to algorithm code.
“As hardware systems become more powerful, MATLAB users are increasingly presented with data-intensive problems that involve highly complex data sets,” says Silvina Grad-Freilich, manager of parallel computing and application deployment marketing at The MathWorks.
In Version 4.2, Parallel Computing Toolbox can now be used with two additional MathWorks toolboxes to accelerate specific algorithms on multiprocessing hardware without requiring users to write or modify a single line of code.
For example, the algorithms in Statistics Toolbox have been modified, including the bootstrap and cross-validation algorithms, which are resampling methods that require repeatedly evaluating statistical functions on multiple data samples.
Also, algorithms in Communications Toolbox have been modified so that you can run computationally intensive simulations of error-rate performance models in parallel. These enhancements build on the existing set of toolbox algorithms that take advantage of parallel operations, such as those in Optimization Toolbox and Genetic Algorithm and Direct Search Toolbox.
To learn more, go to www.mathworks.com.