I guess though I was curious if the GPU has to be on the server itself, or if it's able to harness a remote GPU. Since the database server is likely to be long running it will be expensive to rent GPU enabled hardware for all the time it is on line.
EVA currently does not support remote GPUs. Our ongoing integration [1] of the Ray distributed compute framework [2] into EVA will soon allow us to support remote GPUs. We would love to jointly explore how to best support remote GPUs. Please consider opening an issue with more details on your use case.
[1] https://github.com/georgia-tech-db/eva/blob/master/eva/exper...
[2] https://docs.ray.io/en/latest/cluster/vms/getting-started.ht...