You can prototype and debug applications on the desktop with Parallel Computing Toolbox and easily scale to clusters and clouds with MATLAB Parallel Server and minimal code change. MATLAB Parallel Server supports both interactive and batch workflows to leverage your cluster hardware.
Through MATLAB and Parallel Computing Toolbox, users in your organization can submit jobs to computational resources configured with MATLAB Parallel Server without being concerned about differences in operating systems, environments, and schedulers. MATLAB Parallel Server integrates MATLAB and Simulink with existing scheduler environments at the application layer.
MATLAB Parallel Server Licensing
MATLAB Parallel Server is licensed separately from MATLAB. It is licensed based on the number of MATLAB computational engines (workers) running simultaneously. For each MATLAB computational engine launched by the scheduler, a worker is checked out from the license. License size is determined by the number of workers you need to run simultaneously. The licensing model includes features to support unlimited scaling.
End users are automatically licensed on the cluster for the MathWorks® products they use on the desktop. The cluster requires only a MATLAB Parallel Server license.
MATLAB Parallel Server can be used with a network license manager, or with online licensing, which is convenient for the cloud and for personal clusters.
To assess your license needs, select one of the options below:
Cluster Environment Option |
Effort Required for Set Up |
Initial Set Up for Each New Cluster |
Supported Job Schedulers |
License Manager |
---|---|---|---|---|
MathWorks Cloud Center |
Low |
Set up with web application or in MATLAB |
MATLAB Job Scheduler |
Online licensing |
MathWorks Reference Architecture for Cloud |
Medium |
Set up with templates |
MATLAB Job Scheduler |
Online licensing, |
Dedicated Clusters for MATLAB |
Medium |
Installation and configuration of software and scheduler |
MATLAB Job Scheduler |
Network license manager, |
General Purpose Clusters and Custom Clouds |
High |
Installation of software and integration with existing scheduler |
Supported third-party schedulers (e.g. Slurm), MATLAB Job Scheduler in Kubernetes |
Network license manager, |
Low |
Handled by partner |
MATLAB Job Scheduler, |
Network license manager, |
All licensing options mentioned above will work with any cluster environment options in the table. If you use online licensing, you must link your MathWorks Account to the license.
*Note: If multiple license manager options are listed, the first option is recommended.
If you have any questions regarding set up, contact Support.