How to increase the number of workers up to 72 ?

2 visualizzazioni (ultimi 30 giorni)
kouki mohamed
kouki mohamed il 30 Nov 2018
Modificato: Walter Roberson il 20 Giu 2025
I have checked my 'numcores', here is the result.
>> feature('numcores')
MATLAB detected: 24 physical cores.
MATLAB detected: 24 logical cores.
MATLAB was assigned: 24 logical cores by the OS.
MATLAB is using: 24 logical cores.
ans=24
The problem is: I would like to run my program on a server on 3 nodes (anode = 24 cores), but when I run my matlab code in parallel, matlab displays this message:
'Starting parallel pool(parpool) using the 'local' profile.. connnected to 24 workers'
My question is how to run my progromme using 72 workers.

Risposte (1)

Leepakshi
Leepakshi il 3 Giu 2025
Hi Kouki,
To run your MATLAB program using 72 workers across 3 nodes, try using MATLAB Parallel Server with a properly configured cluster profile rather than the default 'local' profile, which limits you to the cores on your local machine. Cluster profiles connect MATLAB to remote computing clusters managed by schedulers, allowing you to scale computations across multiple nodes and hundreds or thousands of cores. You create or manage these profiles using the Cluster Profile Manager as described in https://www.mathworks.com/help/parallel-computing/discover-clusters-and-use-cluster-profiles.html.
Starting a parallel pool with such a cluster profile via parpool(parcluster('YourProfile'), 72) submits jobs to the cluster scheduler to allocate 72 workers across nodes, enabling distributed parallel processing beyond local hardware limits.
For details on MATLAB Parallel Server and its capabilities, check below link https://www.mathworks.com/products/matlab-parallel-server.html.
Hope this helps!
Thanks
  1 Commento
Raymond Norris
Raymond Norris il 20 Giu 2025
@Leepakshi I'd make one edit to your answer. Change:
parpool(parcluster('YourProfile'), 72)
to
parpool('YourProfile', 72)
parpool can take a cluster object (in your case) or you can pass it the name of the profile to use (in my case)

Accedi per commentare.

Categorie

Scopri di più su MATLAB Parallel Server in Help Center e File Exchange

Community Treasure Hunt

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

Start Hunting!

Translated by