Skip to content

Latest commit

 

History

History
83 lines (57 loc) · 3.75 KB

README.md

File metadata and controls

83 lines (57 loc) · 3.75 KB
Raw

ProActive Matlab/Scilab Connector

Build Status

ProActive Matlab/Scilab connector is part of the ProActive Parallel Scientific Toolbox that aims to accelerate application deployment and execution by distributing and parallelizing executions but also managing data transfers on Desktop machines, Clusters, Grids and Clouds.

Building from sources

You can build from the master branch the bundles that contain the binaries for Matlab and Scilab with gradle build.

This will produce the following archives:

build/distributions/
├── matlab_scilab_connector-XXXXX-SNAPSHOT-matlab.zip
└── matlab_scilab_connector-XXXXX-SNAPSHOT-scilab.zip

Quick start

  1. Follow the installation instructions for ProActive Scheduling.

  2. Unzip the desired distribution archive. The path to this folder will be referred as CONNECTOR_HOME in the following.

  3. Copy the contents of folder CONNECTOR_HOME/lib/worker into SCHEDULING_HOME/addons/client/ folder.

    3a. If you use Matlab, edit the XML file MatlabWorkerConfiguration.xml inside the SCHEDULING_HOME/addons/client/ folder according to your local Matlab installation. The MachineGroup tag allows to specify a range of host for which the given configuration applies. Several configurations for several machine groups can be written in a single MatlabWorkerConfiguration.xml file, but this makes sense only if the scheduler is installed in a shared folder and every worker Node will use the same scheduler installation when starting and registering to the ResourceManager. When each machine uses a local scheduler worker installation, then the content of CONNECTOR_HOME/lib/worker folder must be copied into each addons/client folder of the scheduler worker installation, and the MatlabWorkerConfiguration.xml file must be edited on each machine.

    3b. If you use Scilab, edit the ScilabWorkerConfiguration.xml file, and do the same by considering Scilab.

  4. Start a scheduler instance and note the url at which the scheduler is started, e.g. pnp://localhost:64738.

  5. Load the connector from your scientific environment and connect to the scheduler.

    5a. If you use Matlab:

    • Open Matlab and add path to the ProActive Matlab toolbox which is located on CONNECTOR_HOME, e.g.:
        addpath('CONNECTOR_HOME')
    • Connect to the scheduler by using the function PAconnect, e.g.:
        PAconnect('pnp://localhost:64738')

    • Enter your Scheduler login information, e.g. demo/demo

    • Do a small execution test:

        r=PAsolve(@factorial, 1)

    val = PAwaitFor(r)

    5b. If you use Scilab:

    • build the ProActive Scilab toolbox located on CONNECTOR_HOME, e.g.:
        cd CONNECTOR_HOME
    exec builder.sce
    • load the toolbox into the Scilab environment (at each Scilab restart only the exec loader.sce command will be necessary)
        exec loader.sce
    • Connect to the scheduler by using the function PAconnect, e.g.:
        PAconnect('pnp://localhost:64738')

    • Enter your Scheduler login information, e.g. demo/demo

    • Do a small execution test:

        r=PAsolve('cosh', 1)
    val = PAwaitFor(r)

For further information, please refers to the Matlab toolbox and Scilab toolbox documentations available from within Matlab and Scilab.

Contact

If you have any problems or questions when using ProActive Matlab/Scilab Connector, feel free to contact us at [email protected]