pert tests: each test re-implements standard boiler plate hindering maintenance

[lucbv]

@brian-kelley @e10harvey The issue is related to PRs #1624 #1664 and #1672 all of which touch on boiler plate logic. This issue should be resolved by the introduction of utility functions and macros. Here are two proposed enhancement

1. introduce a struct and functions to handle listing, parsing and setting of common inputs (--cuda, --openmp, etc...)
2. create a macro that performs some logic checks and picks execution spaces on which to run the performance test

Let me know if additional features would be useful, a PR with the above changes will be issued soon

[lucbv]

Maybe we should lift some of the code in kokkos/core/src/impl/Kokkos_Command_Line_Parsing.cpp? : )

[cwpearson]

The other thing worth thinking about is interaction with Google Benchmark, which parses / consumes its own set of CLI arguments (I think they all start with --benchmark_)