How can I implement MATLAB parallel computing on more than one node

61 vues (au cours des 30 derniers jours)
Shuaibin WAN
Shuaibin WAN le 11 Juil 2019
Commenté : Shuaibin WAN le 2 Mar 2021
Hi everyone! I have a problem in implementing MATLAB parallel computing.
On a supercomputing server, I assign the task to 2 nodes (48 workers in maximum).
But the MATLAB command 'parpool(36)' cannot work. The error message is shown in the picture below.
I feel so confused about it and havn't found a solution to it in the community.
I really appreciate it if anyone can help me! Thank you so much!
Capture2.PNG

Réponse acceptée

lomula satewilo
lomula satewilo le 1 Mar 2021
Modifié(e) : lomula satewilo le 2 Mar 2021
As the error message shows, the cluster "local" only allows for a maximum of 24 workers (equal to the number of physical cores on the local node). To recruit workers on more than one node, another cluster must be defined and used:
c = parcluster('My_cluster_profile');
p = parpool(c, 36);
Use Cluster Profile Manager in Matlab to discover or create cluster profiles. For running jobs on multiple nodes, Matlab Parallel Server has to be installed too, in addition to Parallel Toolbox. Otherwise, you can only run parallel jobs on the local machine.
  1 commentaire
Shuaibin WAN
Shuaibin WAN le 2 Mar 2021
Thank you so much for your help!
Yes, you are right. Usually, MATLAB only use local workers.
Later I found local workers were kind of enough for my research, although I had to spend more days to wait for my results lol.

Connectez-vous pour commenter.

Plus de réponses (1)

Aya Ghalayini
Aya Ghalayini le 14 Août 2020
I had the same problem and I was told by my administrator that Matlab seems to use only one node. See this link as well: https://stackoverflow.com/questions/17978421/matlab-parallel-processing-on-several-nodes
I hope this helps.

Catégories

En savoir plus sur Parallel Computing Fundamentals dans Help Center et File Exchange

Produits


Version

R2017b

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by