Table of Contents
Introduction
There are several ways in which to submit MATLAB jobs to a cluster. This document will cover the various ways to run MATLAB compute jobs on the Shared Research Compute clusters, which will include using the Parallel Computing Toolbox (PCT) and the MATLAB Distributed Compute Server (MDCS) to submit many independent tasks as well as to submit a single task that has parallel components. Examples are included.
Definitions
Task Parallel Application - The same application that runs independently on several nodes, possibly with different input parameters. There is no communication, shared data, or synchronization points between the nodes.
Data Parallel Application - The same application that runs on several labs simultaneously, with communication, shared data, or synchronization points between the labs.
Lab - A MATLAB worker in a multicore (Data Parallel) job. One lab is assigned to one worker (core). Thus, a job with eight labs has eight processor cores allocated to it and will have eight workers each working together as peers.
MDCS - Matlab Distributed Compute Server. This is a component of MATLAB that allows our clusters to run MATLAB jobs that exceed the size of a single compute node (multinode parallel jobs). It also allows jobs to run even if there are not enough toolbox licenses available for a particular toolbox, so long as the university owns at least one license for the particular toolbox.
PCT - Parallel Computing Toolbox.
MATLAB Task - One segment of a job to be evaluated by a worker.
MATLAB Job - The complete large-scale operation to perform in MATLAB, composed of a set of tasks.
MATLAB Worker - The MATLAB session that performs the task computations. If a job needs eight processor cores, then it must have eight workers.
Job - Job submitted via the PBS job scheduler (also called PBS Job).
Interactive Jobs
TBA take from old FAQ
Running Jobs with PBS qsub
The recommended method of submitting Task Parallel jobs is to use PCT as described below. However, if you need to submit one or more single core jobs, and are not encountering problems with the lack of specific toolbox licenses, then you might wish to submit your jobs directly with PBS as described in the following steps.
1. Build a MATLAB .m code file
Include all of the MATLAB commands that you need to execute in a MATLAB .m file, such as sample.m and place it somewhere in your home directory or subdirectory. The creation and contents of a .m file are beyond the scope of this document. Consult the MATLAB documentation for information on .m files.
2. Create a PBS batch script
You will need to execute Matlab from within a PBS batch script, such as sample.pbs as follows. In this example, this file is saved in the same directory as sample.m.
In this example the sample.pbs script calls matlab with the -r option followed by the MATLAB script name that was created in step #1. Leave off the trailing .m from the script name when callingMATLAB this way.
For more information about PBS job scripts, please see our FAQ.
3. Submit the Job
After you have created sample.m and sample.pbs, go to the directory where sample.pbs resides, load the MATLAB module (if you have not already done so) and submit the job to the job scheduler:
Submitting multiple MATLB jobs via PBS
In order to submit multiple Matlab batch jobs, simply repeat this section for each job.
Using PCT for Task Parallel and Data Parallel Jobs
PCT provides an API that allows you to submit a job that has multiple independent tasks (Task Parallel) or submit a job that has a single task that is a multiprocessor, and possibly multinode, task (Data Parallel). In order to run this type of job you must first configure MATLAB for this type of job submission by following these steps;
Configuring MATLAB
1. In your home directory create the MdcsDataLocation subdirectory.
2. Load the MATLAB 2011a environment:
3. Run MATLAB on the login node:
4. In MATLAB, add the /opt/apps/matlab/2011a-scripts folder to your MATLAB path so that MATLAB will be able to find the scripts necessary to submit and schedule jobs.
- Click File > Set Path
- Click Add Folder
- Specify the following folder:
/opt/apps/matlab/2011a-scripts
5. Import the cluster configuration for the cluster you are running on:
- On the MATLAB Desktop menu bar, click Parallel > Manage Configurations.
- Click File > Import
- In the Import Configuration dialog box, browse to find the MAT-file for the configuration that you want to import. Navigate to /opt/apps/matlab/2011a-scripts and select the configuration for the system you are using, such as sugar.mat, davinci.mat, stic.mat, and so forth. Select the file and click Import.
- Select the configuration you have imported and click Start Validation.
- All four stages should pass: Find Resources, Distributed Job, Parallel Job, Matlabpool
If all validation stages succeed, then you are ready to submit jobs to MDCS.
Submitting Task Parallel Jobs
The following is an example of a Task Parallel job. The task-parallel example code, frontDemo, calculates the risk and return based on historical data from a collection of stock prices. The core of the code, calcFrontier, minimizes the equations for a set of returns. In order to parallelize the code, the for loop is converted into a parfor loop with each iteration of the loop becoming its own independent task.
To submit the job, copy submitParJobToCluster.m into your working directory, make the necessary modifications for your job environment, and then run the code from within MATLAB. This will submit the job. An explanation of the code follows:
Code
When you run this code within MATLAB, the frontDemo code will be submitted to the PBS job scheduler. Use the showq command from a cluster terminal window to look for your job in the job queue.
Submitting Data Parallel Jobs
The data-parallel example code calculates the area of pi under the curve. The non parallel version, calcPiSerial, calculates with a for loop, looping through discrete points. The parallel version, calcPiSpmd, uses the spmd construct to evaluate a part of the curve on each MATLAB instance. Each MATLAB instances uses its labindex (i.e. rank) to determine which portion of the curve to calculate. The calculations are then globally summed together and broadcasted back out. The code uses higher level routines, rather than lower level MPI calls. Once the summation has been calculated, it’s indexed into and communicated back to the local client MATLAB to calculate the total area.
To submit the job, copy submitSpmdJobToCluster.m into your working directory, make the necessary modifications for your job environment, and then run the code from within MATLAB. This will submit the job. An explanation of the code follows:
When you run this code within MATLAB, the calcPiSpmd code will be submitted to the PBS job scheduler. Use the showq command from a cluster terminal window to look for your job in the job queue.
Code
Job Dependencies
In order to run code on the cluster, a job may be dependent on several MATLAB or data files. The batch() function takes two parameters: PathDependencies and FileDependencies. Both can be assigned to a comma separated cell array of filenames and/or folder names. If the MATLAB client shares a file system with the compute nodes, then typically the user will specify the dependencies on the local path (i.e PathDependencies). For example:
Configuring Cluster Parameters with ClusterInfo
Destroying a Job
When you submit a job with batch, you will notice that each submission is labeled Job1, Job2, and so forth. Temporary directories associated with each job can be found in ~/MdcsDataLocation as the jobs are running. When job.destroy is called, these temporary directories are deleted. The above examples call job.destroy. If you close your MATLAB session before executing job.destroy, which is likely unless you are using the full example with job.wait, you will need to manually cleanup temporary directories in ~/MdcsDataLocation.
Running Locally on a Desktop
In order to run MATLAB code with a parallel component on your desktop locally, you must first start up a MATLAB Pool, as such:
where 8 is the number of MATLAB processes to attach to the job. At this point you will have access to eight MATLAB workers for use with parallel code, such as code with parfor loops, and so on.
After running the code, close the MATLAB Pool: