Control over GPU <-> CPU result copies?

[lumost]

is it possible to control whether the results from a given GPU search are CPU/Host memory resident or GPU memory resident?

[Enet4]

This is an intriguing question. These bindings are implemented through the C API, which in turn uses the native C++ API. To the best of my knowledge, the high level API here does not provide that sort of awareness.

If you can provide an example of how this would be done in C++ through the main Faiss API, then we may have a clearer idea of how this could be achieved in Rust.