A user wishes to use math application running on our PBSPpro cluster.

Our cluster is accessible to linux users, running qsub to submit their jobs.

We have math10.4 for our cluster users, with infinite kernel licenses (we are about to upgrade to 11)

The user has standalone 10.4 math on his mac PC, and he would like to utilize our cluster's cores in addition to his PC's

Reading some internet tutorials and examples, I have managed to setup the math application on his PC, and also on my PC (windows 7 standalone math). I succeeded to send commands to our remote head node, from my math application:

I also managed to run 6 math kernels on our head node (which has 6 cores) enter image description here After executing this command, I see in our cluster head node 6 Wolfram processes.

However, the user's goal is to run calculations in parallel utilizing our cluster workers - nodes and cores (700-800 cores), and if I set more than 6 (e.g. 24 kernels) – like in the below config – then seems that the allocated kernels are all on our head node, not using the working nodes. enter image description here

Can someone please help me with a solution of utilizing both PC and cluster cores?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.