Awesome Open Source
Awesome Open Source


Parallel computing without the Parallel Computing Toolbox

This is basically a "fork" of, with some convenience functions added for my own convenience.

All credit goes to Markus Buehren

See also


  • added spawnmulticoreslaveinstances and killmulticoreslaveinstances functions to automate instance launching
  • made code usable as a package (simply copy files to path\+multicore\)

Basic usage

The simplest usage example is:

results = multicore.startmulticoremaster(@functionHandle, parameterCell);

This will

  • determine the number of available (physical) cores on the local machine
  • start slave instances such that the total number of workers (including master) matches the number of cores
  • perform the calculations
  • remove all temp files
  • shut down all slaves

Advanced usage

multicoreDir = multicore.spawnmulticoreslaveinstances(n_slaves, multicoreDir, settings)

allows to pass the same settings to each of n_slaves, using multicoreDir for the job pool.

Additional machines can be pointed at the same multicoreDir.


will shut down all slaves working on the job pool in multicoreDir

Get A Weekly Email With Trending Projects For These Topics
No Spam. Unsubscribe easily at any time.
Matlab (44,777
Parallel Computing (690
Parallel Processing (158
Related Projects