I have two compute clusters of V100 GPU each(named as - "Cluster1" with 2 nodes and "Cluster2" with 2 nodes), I want to use both these clusters in my training script(Pytorch Training).
Right now, I can use "Cluster1" or "Cluster2" in my Compute target code and my training code is getting 1 V100 GPU for training. Can anybody help on how to use both Clusters(Cluster1 and Cluster2) in my Compute target code, so that my training script can see 2 V100 GPUs not 1 GPU.
Can you tell how this will be achieved, also please make me understand what are the 2 nodes means in one V100 GPU cluster(it is not two GPUs that I know for sure). But then what does that 2 nodes in one cluster means and how it works??