https://github.com/simonster/matlabcluster.jl

Julia cluster manager for MATLAB Job Scheduler
https://github.com/simonster/matlabcluster.jl

Last synced: about 1 month ago
JSON representation

Julia cluster manager for MATLAB Job Scheduler

• Host: GitHub
• URL: https://github.com/simonster/matlabcluster.jl
• Owner: simonster
• License: other
• Created: 2013-10-11T04:18:18.000Z (about 11 years ago)
• Default Branch: master
• Last Pushed: 2014-10-16T03:39:36.000Z (about 10 years ago)
• Last Synced: 2024-10-13T19:33:06.053Z (2 months ago)
• Language: Julia
• Homepage:
• Size: 188 KB
• Stars: 1
• Watchers: 2
• Forks: 4
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE.md

Awesome Lists containing this project

README

        
# MATLABCluster.jl

The `MATLABCluster.jl` package provides an interface for starting [Julia](http://www.julialang.org/) worker processes using [MATLAB™ Parallel Computing Toolbox](http://www.mathworks.com/help/distcomp/index.html) cluster profiles, including profiles configured to connect to a [MATLAB™ Distributed Computing Server](http://www.mathworks.com/products/distriben/index.html). You cannot use `MATLABCluster.jl` without having purchased and installed a copy of MATLAB™ and the MATLAB™ Parallel Computing Toolbox from [MathWorks](http://www.mathworks.com/). This package is available free of charge and in no way replaces or alters any functionality of MathWorks's MATLAB product.

To connect to a MATLAB cluster:

```julia

using MATLABCluster

addprocs(1,                               # The number of workers you want

         dir="/path/to/julia/usr/bin",    # Path to Julia on your cluster

         cman=MATLABManager("profile"),   # The cluster profile, set up within MATLAB

         tunnel=true,                     # Only necessary if the workers are behind NAT

         sshflags=`-c blowfish`           # Only for NAT; faster than AES (default)

        )

```

## Caveats

- Currently output from the workers is not transmitted back to Julia

- This may not work if the MATLAB Job Scheduler is running on Windows (but let me know if you try)