Some jobs require more control than the functionality offered
by high-level constructs like spmd and parfor.
In such cases, you have to program all the steps for creating and
running the job. Using the local cluster (or local scheduler) on your
machine lets you create and test your jobs without using the resources
of your network cluster. Distributing tasks to workers that are all
running on your client machine do not offer any performance enhancement.
Therefore this feature is provided primarily for code development,
testing, and debugging.
Workers running in a local cluster on a Microsoft® Windows® operating
system can display Simulink® graphics and the output from certain
functions such as uigetfile and uigetdir. (With other platforms or schedulers,
workers cannot display any graphical output.) This behavior is subject
to removal in a future release.
This section details the steps of a typical programming session with Parallel Computing Toolbox™ software using a local cluster:
The objects used by the client session to interact with the
cluster are only references to data in the cluster job storage location,
not in the client session. After jobs and tasks are created, you can
close your client session and restart it, and your job still resides
in the storage location. You can find existing jobs using the findJob function or the Jobs property
of the cluster object.
You use the parcluster function
to create an object in your local MATLAB® session representing
the local scheduler.
parallel.defaultClusterProfile('local');
c = parcluster();You create a job with the createJob function.
This statement creates a job in the cluster job storage location and
creates the job object job1in the client session.
If you omit the semicolon at the end of the command, it displays some
information about the job.
job1 = createJob(c)
Job
Properties:
ID: 2
Type: Independent
Username: eng864
State: pending
SubmitTime:
StartTime:
Running Duration: 0 days 0h 0m 0s
AutoAttachFiles: true
Auto Attached Files: List files
AttachedFiles: {}
AdditionalPaths: {}
Associated Tasks:
Number Pending: 0
Number Running: 0
Number Finished: 0
Task ID of Errors: []TheState property of the job is pending.
This means that the job has not yet been submitted (queued) for running,
so you can now add tasks to it.
The scheduler display now indicates the existence of your job, which is the pending one, as appears in this partial listing:
c
Local Cluster
Associated Jobs
Number Pending: 1
Number Queued: 0
Number Running: 0
Number Finished: 0After you have created your job, you can create tasks for the
job using the createTask function.
Tasks define the functions to be evaluated by the workers during the
running of the job. Often, the tasks of a job are all identical. In
this example, five tasks each generate a 3-by-3 matrix of random numbers.
createTask(job1, @rand, 1, {{3,3} {3,3} {3,3} {3,3} {3,3}});The Tasks property of job1 is
now a 5-by-1 matrix of task objects.
job1.Tasks
ID State FinishTime Function Error
-----------------------------------------------------
1 1 pending @rand
2 2 pending @rand
3 3 pending @rand
4 4 pending @rand
5 5 pending @rand To run your job and have its tasks evaluated, you submit the
job to the cluster with the submit function.
submit(job1)
The local scheduler starts the workers on your machine, and
distributes the tasks of job1 to these workers
for evaluation.
The results of each task evaluation are stored in the task object OutputArguments property
as a cell array. After waiting for the job to complete, use the function fetchOutputs to retrieve the results
from all the tasks in the job.
wait(job1) results = fetchOutputs(job1);
Display the results from each task.
results{1:5} 0.9501 0.4860 0.4565
0.2311 0.8913 0.0185
0.6068 0.7621 0.8214
0.4447 0.9218 0.4057
0.6154 0.7382 0.9355
0.7919 0.1763 0.9169
0.4103 0.3529 0.1389
0.8936 0.8132 0.2028
0.0579 0.0099 0.1987
0.6038 0.0153 0.9318
0.2722 0.7468 0.4660
0.1988 0.4451 0.4186
0.8462 0.6721 0.6813
0.5252 0.8381 0.3795
0.2026 0.0196 0.8318After the job is complete, you can repeat the commands to examine the updated status of the cluster, job, and task objects:
c job1 job1.Tasks
The local scheduler runs in the MATLAB client session,
so you do not have to start any separate scheduler or MJS process
for the local scheduler. When you submit a job to the local cluster,
the scheduler starts a MATLAB worker for each task in the job.
You can do this for as many workers as allowed by the local profile.
If your job has more tasks than allowed workers, the scheduler waits
for one of the current tasks to complete before starting another MATLAB worker
to evaluate the next task. You can modify the number of allowed workers
in the local cluster profile. If not specified,
the default is to run only as many workers as computational cores
on the machine.
The local cluster has no interaction with any other scheduler or MJS, nor with any other workers that can also be running on your client machine under the mdcs service. Multiple MATLAB sessions on your computer can each start its own local scheduler with its own workers, but these groups do not interact with each other.
When you end your MATLAB client session, its local scheduler and any workers that happen to be running also stop immediately.