Unable to build with old CUDA version (`CUDA_COMPUTE_CAP = 52`)

[nikvoid]

Hi, I'm unable to build this crate without changes for old GPU (Tesla M40, capability level 5.2, driver 462). Most of the problems are related to FP16 and the fact that FP16 functions are disabled only in Rust code, but no changes are made to CUDA code. Also there was a minor issue with Nvidia-SMI's --query-gpu=compute_cap, so I have to comment out that code and specify capability level manually.

I've made few dirty patches to make it work, in particular that's comment out Nvidia-SMI query, #define __half float, then comment out FP16 functions with compilation failures.