This is machine translation

Translated by Microsoft
Mouse over text to see original. Click the button below to return to the English verison of the page.


Configure options for MPI communication




mpiSettings('DeadlockDetection','on') turns on deadlock detection during calls to labSend and labReceive. If deadlock is detected, a call to labReceive might cause an error. Although it is not necessary to enable deadlock detection on all workers, this is the most useful option. The default value is 'off' for communicating jobs, and 'on' inside pmode sessions or spmd statements. Once the setting has been changed within a pmode session or an spmd statement, the setting stays in effect until either the pmode session ends or the parallel pool is closed.

mpiSettings('MessageLogging','on') turns on MPI message logging. The default is 'off'. The default destination is the MATLAB® Command Window.

mpiSettings('MessageLoggingDestination','CommandWindow') sends MPI logging information to the MATLAB Command Window. If the task within a communicating job is set to capture Command Window output, the MPI logging information will be present in the task's CommandWindowOutput property.

mpiSettings('MessageLoggingDestination','stdout') sends MPI logging information to the standard output for the MATLAB process. If you are using a MATLAB job scheduler (MJS), this is the mdce service log file.

mpiSettings('MessageLoggingDestination','File','filename') sends MPI logging information to the specified file.


Set deadlock detection for a communicating job inside the jobStartup.m file for that job:

    % Inside jobStartup.m for the communicating job
    myLogFname = sprintf('%s_%d.log',tempname,labindex);

Turn off deadlock detection for all subsequent spmd statements that use the same parallel pool:


More About

collapse all


Setting the MessageLoggingDestination does not automatically enable message logging. A separate call is required to enable message logging.

mpiSettings has to be called on the worker, not the client. That is, it should be called within the task function, within jobStartup.m, or within taskStartup.m.

Was this topic helpful?