Skip to content

Latest commit

 

History

History
83 lines (56 loc) · 2.7 KB

README.md

File metadata and controls

83 lines (56 loc) · 2.7 KB
Raw

A) About

This is version 0.2.7 of ParforProgress2, a simple parfor progress monitor for matlab. See also http://www.mathworks.com/matlabcentral/fileexchange/35609-matlab-parforprogress2.

This progress monitor comes with a nice wrapper ParforProgressStarter2.m which will take care of adding the classes to the java class path, depending on whether matlabpool / parpool is enabled or not.

B) Setup

  • clone this directory
  • add directory to matlab path: addpath('/path/to/matlab-ParforProgress2')

C) How to use

By default ParforProgress2 is very easy to use:

  1. add program location to matlab path
  2. run ParforProgress2.
    See C1 - no globals below for details.

In case your Matlab programs make use of global variables, things are a bit more complicated. See C2 - with globals below for details.

C1) no globals

See ParforProgressStressTest2.m or ParforProgressStarter2.m for usage instructions.

C2) with globals

ParforProgress2 makes use of javaaddpath and javaclasspath. Using either of these functions makes Matlab call clear java which does clear all global variables. This Matlab behaviour is intentional. See http://www.mathworks.com/help/matlab/matlab_external/bringing-java-classes-and-methods-into-matlab-workspace.html for details.

In order to make ParforProgress2 work with globals, see the following instructions, which depend on your used Matlab version.

C2.1) Matlab 2012b and later:

  • create javaclasspath.txt file in your preference directory.
  • add path to ParforProgressStarter2 to this file.

On Linux this translates to:

$ cd .matlab/R2013a
$ echo "/path/to/matlab-ParforProgress2" > javaclasspath.txt

Next, startup Matlab and call ParforProgressStarter2, but do not run javaddpath:

 >> % setup parameters
 >> run_javaaddpath = 0;
 >> s = 'dummy task';
 >> n = 100;
 >> percentage = 0.1;
 >> do_debug = 0;
 >>
 >> % initialize the ProgressMonitor
 >> ppm = ParforProgressStarter2(s, n, percentage, do_debug, run_javaaddpath)
 >> 
 >> % run your computation
 >> for j = 1 : n
 >>     your_computation();
 >>     ppm.increment(i); 
 >> end
 >>
 >> % delete the ProgressMonitor
 >> delete(ppm);

C2.2) Matlab 2012a and earlier:

Unfortunately Matlab 2012a and earlier versions of Matlab do not support customized javaclasspath.txt files, instead, you need to edit the classpath.txt file that is part of the Matlab installation. You might need administrator / superuser privileges to edit this file. Here's how you can locate it:

>> which classpath.txt -all
/share/pkgs/MATLAB/R2012a/toolbox/local/classpath.txt
>> edit classpath.txt