MathWorks parallel computing products provide high-level tools that let your users solve large-scale problems using MATLAB and Simulink. With little training, users can take advantage of available high-performance computing resources without leaving their development environment.
MathWorks parallel computing products directly support commercially available resource managers or schedulers, such as Windows® HPC Server, Windows Compute Cluster Server, Platform LSF®, PBS, and TORQUE. The parallel computing products also provide an API and sample scripts for integrating other schedulers, such as SGE, Grid MP™, or DataSynapse.
Parallel computing products support most platforms available on general-purpose clusters—Linux®, Windows, and Mac®—and require no specialized hardware. The products work alongside other applications in your environment, reducing your software management workload.
Users with Parallel Computing Toolbox can run MATLAB workers locally to prototype and test applications on their multicore desktop, freeing shared cluster resources for other users to run their applications.
A floating license scheme for the server offers a cost-effective option and simplifies the management of cluster software. The scheme requires a single cluster license for the MATLAB Distributed Computing Server and lets multiple users with different products on their desktops access the server without requiring additional individual licenses to be managed on the cluster computers.
"It took us about an hour to move our simulations from a single processor to a distributed, 16-node cluster."Read the story