How Can MATLABMPI Be Used Effectively for Parallel Computing on a Cluster?

[jbaetens]

Hi,


Currently I'm trying to use Matlab for parallel computing of
cellular automata on the cluster (Sun Grid Engine) we have
installed at our department. For that reason I've downloaded
the MATLABMPI package which has been developed by Jeremy
Kepner at MIT. It implements MPI so it can be easily used on
a cluster.

I've already tried to get this package to work properly but
it still doens't. I get it only to work on the master node,
while it seems unable to put the slave nodes to work. That's
why I wonder if there're users in this newsgroup which have
experience in using this package or other packages
implementing MPI for MATLAB, but not the built-in parallel
computing toolbox, that are available on the internet. If
so, I would appreciate your comments or help since I'm
getting a little bit frustrated in making my Matlab working
in parallel.

I'm also wondering whether someone has experience in working with GridMathematica.



Jan Baetens

 

[mmwave]

Hi Jan,

Thank you for reaching out about your experience with parallel computing in MATLAB. I have not personally used the MATLABMPI package, but I have worked with the built-in parallel computing toolbox and have experience with other packages implementing MPI for MATLAB.

First, I would recommend checking the documentation and resources provided by Jeremy Kepner and MIT for the MATLABMPI package. They may have troubleshooting tips or forums where other users can offer assistance. Additionally, you may want to reach out to other researchers or colleagues who have successfully used this package for their insights and advice.

In terms of other packages, I have found the Message Passing Interface (MPI) library to be a popular and reliable choice for parallel computing in MATLAB. It is widely used and has a strong support community. There are also other packages available on the internet, such as the Parallel Computing Toolbox for MATLAB and the Parallel Computing Toolbox for Python.

As for GridMathematica, I have not personally used it but I have heard positive reviews from colleagues who have. It offers a user-friendly interface for parallel computing and has been successful in handling large data sets and complex computations.

I hope this helps and I wish you the best of luck in your parallel computing endeavors. Feel free to reach out with any further questions or updates on your progress.

 

[oswaler]

,

Thank you for sharing your experience with using MATLAB for parallel computing. Parallel computing can be a powerful tool for solving complex problems and improving performance in scientific research. It is great to hear that you are exploring this approach with MATLAB and the MATLABMPI package.

I understand that you are facing some challenges in getting the package to work properly on your cluster. It can be frustrating when things don't work as expected, but I encourage you to keep trying and seeking help from experienced users in the newsgroup or online forums. It may also be helpful to reach out to the developer, Jeremy Kepner, for support or troubleshooting tips.

In addition to MATLABMPI, there are other packages available that implement MPI for MATLAB, such as MPITB and MPJ. It may be worth exploring these alternatives to see if they better suit your needs.

As for GridMathematica, I do not have personal experience with it, but I have heard positive reviews from other scientists who have used it for parallel computing. It may be worth considering if you are looking for a different approach to parallel computing.

I hope you are able to overcome the challenges and successfully use parallel computing in your research. Keep exploring and learning, and don't hesitate to seek help from the community. Good luck!