Training neural network using parallel computing on high-performance computing (HPC) cluster.

I am trying to train a neural network on a high-performance computing (HPC) cluster. When I use the option of parallel computing on my computer Matlab uses all the available cores to train the network. When I have many more cores in the cluster, I see the program is using only one core despite using parallel options for training a neural network. Can anyone know how to train a neural network on Matlab using multiple cores of a high-performance computing (HPC) cluster?

Answer 1

[David Willingham]

Hi Pardhasai,

MATLAB Parallel Server enables you to access all the cores on a cluster. It handles all the back end scheduling and communication between the nodes on the cluser for you. Once setup, you simply select it from your Parallel Prefences in MATLAB.

Regards,

David Willingham

Deep Learning Product Manager