Add more info to ANN_BENCH context

[achirkin]

Add extra information to benchmark context for better reproducibility and performance analysis:

1. Full command line used to call the executable (so you can copy-paste and run again).
2. More CUDA device information: whether HMM, AST, or host atomics are available (how GPU can efficiently communicate with CPU).
3. Host information: min/max frequences, used virtual processors and cores, available physical memory and swap (does the benchmark segfault due to not enough host memory? is SMT enabled? etc).

Addresses parts of #160

[achirkin]

I'm thinking whether we should include some environment variables if present or the machine hostname? Wouldn't that be a bit fishy in terms of user privacy when sharing the output files?

[tfeher]

Thanks Artem, LGTM, it is useful to extend the context information. I do not see any issues with adding the hostname.

[tfeher]

For env vars, I am not sure what would be relevant.
For OpenMP context, you could add omp_get_max_threads()

[achirkin]

We could get the openmp-related env vars to see if the limit to the number of threads is set explicitly.
I wouldn't like to add omp_get_max_threads() and other OpenMP functions to avoid introducing openmp dependency in the benchmark (for the algorithms that do not use OpenMP and for the ANN_BENCH executable in the single-exe mode). I'm also not sure if cuVS team decides to move away from using OpenMP in favor of standard C++ threads in future.

[tfeher]

/merge