Add more C backend stuff accessible to python

*) compute capability from xgboost code
*) deviceQuery
*) malloc (for testing)
https://stackoverflow.com/questions/5730010/pycuda-querying-device-status-memory-specifically

Start by reviewing the existing C/C++ backend and its Python access points, then compare the requested compute capability, deviceQuery, and malloc items with the linked Stack Overflow discussion. Done means the requested GPU information and test allocation are accessible from Python, with the scope and validation approach clarified.

Written by the indexing model from the issue text.