Multiple Nvidia Titan RTX cluster

I am finding a way to create a multiple GPU cluster using multiple Titan RTX located on different machines within same network.
Also is it possible to get information of all GPUs using ‘nvidia-smi’ on single command ?
If yes then kindly provide the solution.
Following are the deatils of my system :
Operating system - Linux (Ubuntu distribution)
GPU - Titan RTX

no, nvidia-smi can only query or get information on the GPUs that are in the machine that it is being run on. It does not work across a network.

So how can I create a cluster using multiple GPU of multiple machines ?
Also which commands can help me ?

It doesn’t require anything special. You can create a cluster using multiple GPU machines the same way you would create a cluster using non-GPU machines. Quite simply, use a network.

And when it comes to running software on that cluster, use a software tool to distribute the workload, for example MPI.

Which cluster would you like to suggest ?