Matlab 2016 parpool only connects to 1 worker

8 visualizzazioni (ultimi 30 giorni)
Saumya
Saumya il 9 Mag 2017
Risposto: Khadija Bouaita il 4 Dic 2022
My machine has 8 cores. I have tried changing the number of workers in parallel preferences, tried making a new local profile, tried parpool(8); none of these seem to work. I have Matlab 9.0.0.341360 (R2016a). My machine has windows 10. I tried parpool on a different machine with 16 cores and Matlab 2016a, encountered the same problem. On this machine, the validation step of a new cluster connected to 16 workers in job test and SPMD job test, but back to 1 worker in actual parpool. I checked bug reports, and this does not seem to be a known bug. What can I do to use the maximum number of workers?
  1 Commento
Edric Ellis
Edric Ellis il 10 Mag 2017
This is definitely not expected. Please contact MathWorks support to resolve this.

Accedi per commentare.

Risposte (2)

Harsh
Harsh il 12 Mag 2017
Make sure the Parallel Computing Toolbox is installed and licensed for use by running the following commands: https://www.mathworks.com/matlabcentral/answers/225996-how-can-i-tell-if-the-parallel-computing-toolbox-installed#answer_264780
I assume you have run the cluster validation for the "local" profile and have the "Preferred number of workers in a parallel pool" property set to >= number of cores on your machine. Instructions for this can be found here:
Run the following command and make sure the machine has the expected number of cores:
>> feature('numcores')
You can also manually start a parallel pool and configure the number of workers. For example:
>> obj = parcluster
>> obj.NumWorkers = 2
>> parpool(2)
Having verified all the above, if the issue still persists, please contact technical support here: https://www.mathworks.com/support/contact_us/
Please be sure to provide a detailed description of the issue and attach any relevant files / code / examples required to investigate the issue.

Khadija Bouaita
Khadija Bouaita il 4 Dic 2022
x=23

Categorie

Scopri di più su Parallel Computing Fundamentals 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