Store 
Parallel Computing Toolbox™ 3.3
Level of Support for Schedulers
MathWorks™ Job Manager
The MathWorks job manager is a simple scheduler that ships with MATLAB® Distributed Computing Server™. The job manager is intended primarily for groups working with small- to medium-sized clusters.
By default, the job manager runs jobs in the order in which they are submitted. It can run on any machine on the network and can accept jobs from multiple Parallel Computing Toolbox™ users. It lets you dynamically add tasks to running jobs and supports callbacks that execute whenever an event, such as a worker finishing a task, occurs.
The job manager coordinates the execution of MATLAB® applications only. The job manager does not provide any security mechanisms; for example, it does not prevent users from accessing or modifying other users' jobs. For advanced resource and security management features, you must use a third-party scheduler.
Directly Supported Third-Party Schedulers
Parallel Computing Toolbox and MATLAB Distributed Computing Server directly support the following third-party schedulers:
No additional integration steps are required for you to use the toolbox and the engine with these schedulers.
Platform LSF, Windows CCS, PBS Pro, and TORQUE provide several advanced capabilities for resource management and security mechanisms, and also allow you to use your cluster for applications other than distributed applications developed in MATLAB.
Callbacks and dynamic addition of tasks to running jobs are not supported for the third-party schedulers.
With the MathWorks job manager, MATLAB Distributed Computing Server workers on the cluster run as a service and remain running at all times. With these schedulers, the workers run as an application that is started for the evaluation of tasks, and stopped when tasks are complete. As a result, you incur an extra overhead of worker start-up and stop time. Create your tasks taking this overhead into account; for small tasks, this may represent a significant percentage of the overall execution time.
Other Third-Party Schedulers
For third-party schedulers that are not directly supported, such as Sun™ Grid Engine, and Grid MP™ (United Devices), Parallel Computing Toolbox and MATLAB Distributed Computing Server provide a generic scheduler interface that lets you integrate the toolbox and the engine in your existing technical computing environment.
The following resources provide information on using third-party schedulers:
- Documentation provides details on integrating with and using third-party schedulers in Using the Generic Scheduler Interface.
- Sample scripts for  integrating with third-party schedulers such as PBS and SSH are available in the following directory of your MATLAB installation:
$MATLABROOT/toolbox/distcomp/examples/integration - Online technical solutions explain the steps required for integrating MATLAB distributed computing products with:
- Sun Grid Engine
- Grid MP (hosted on MATLAB Central)
- Other schedulers
Callbacks and dynamic addition of tasks to running jobs are not supported for the third-party schedulers.
With the MathWorks job manager, MATLAB Distributed Computing Server workers on the cluster run as a service and remain running at all times. With these schedulers, the workers run as an application that is started for the evaluation of tasks, and stopped when tasks are complete. As a result, you incur an extra overhead of worker start-up and stop time. Create your tasks taking this overhead into account; for small tasks, this may represent a significant percentage of the overall execution time.
