Hint mpirun to pin to one thread per physical core

[nliw]

Hi all. I used to manually set I_MPI_PIN_PROCESSOR_LIST on SMT-enabled nodes. But if the available nodes are of different architecture, doing so by hand will be a real pain. Are there any automatic methods to generate this env by detecting hardware topology? Or achieve the same pinning (one rank per physical core) by setting other envs.

I_MPI_PIN_DOMAIN seems the most close one, except it binds one rank to two logical cores with SMT on.

[VarshaS_Intel]

Hi,

Thanks for posting in Intel Communities.

Please find the environment variable "I_MPI_PIN_PROCESSOR_LIST" for one-to-one pinning. Please find the below link for more information:

https://www.intel.com/content/www/us/en/develop/documentation/mpi-developer-reference-linux/top/environment-variable-reference/process-pinning/environment-variables-for-process-pinning.html

Thanks & Regards,

Varsha

[VarshaS_Intel]

Hi,

We have not heard back from you. Could you please provide an update on your issue?

Thanks & Regards,

Varsha

[VarshaS_Intel]

Hi,

We have not heard back from you. This thread will no longer be monitored by Intel. If you need further assistance, please post a new question.

Thanks & Regards,

Varsha