Inquiry About OpenMPI Support for NVSwitch in Inter-Node GPU Communication

[JieRen98]

Greetings,

I have recently been learning about the CUDA-aware capabilities of OpenMPI and found that it supports GPU peer-to-peer memory access via NVLink within a single node. I am curious to know whether OpenMPI extends this support to inter-node peer-to-peer memory access via NVSwitch.

I would greatly appreciate any insights you could provide on this matter.

Thank you for your time and assistance.

Best,
Jie