To simplify development of multiprocessor and
distributed computing applications, Mathworks has released its Distributed Computing Toolbox 2, which has added support for third-party schedulers, and new
interprocess communication capabilities for distributing and executing parallel algorithms in a cluster of computers.
With support for third-party schedulers such as LSF from Platform Computing, the toolbox, a part of the MATLAB distributed computing portfolio, allows developers to use the generic API provided with the
toolbox and integrate MathWorks distributed computing tools into their existing distributed computing environments.
According to Lisa Kempler, director of MATLAB product marketing, The MathWorks, the toolbox allows developers to take advantage of the unique capabilities of the scheduler, such as support for batch jobs, in addition to the interactive workflows supported in version 1 by The MathWorks job
manager in the MATLAB Distributed Computing Engine.
A major enhancement in the Distributed Computing Toolbox 2 is interprocess communication, enabling execution of parallel applications that are divided into interdependent tasks. The new version includes
communication functions based on Message Passing Interface (MPI), the industry-wide protocol for communication in a parallel program.
Additionally, because Distributed Computing Toolbox 2 runs on all hardware on which MATLAB runs, users can redeploy their parallel applications on new hardware or operating systems without having to retool the entire application.
The Distributed Computing Toolbox supports the full MATLAB language, almost all MathWorks products and all supported MATLAB platforms.
The MATLAB Distributed Computing Engine can run in either homogeneous or heterogeneous clusters. Both the Distributed Computing Toolbox and MATLAB Distributed Computing Engine are available now with prices starting at $1,000 US for the toolbox, and
$6,000 for the engine.
The MathWorks, Inc.