Backports to C++11 of <bit>

[wmaxey]

<cuda/std/bit> - C++20 backport of bit manipulation intrinsics

Allows manipulation of unsigned values using <bit>

Details

Motivation

• CUDA has intrinsics for popc/clz/ctz but they differ from host compiler builtins. Libcu++ has an opportunity to unify host and device code for these intrinsics.
• int128 in device code is not able to make use of CUDA intrinsics, this backport provides a general method for users to manipulate wider values.

Impact

• Backports of to C++11 requires that most functions are minimized into one statement, this impacts readability of the header.
• Conditional use of is_constant_evaluated is required to improve performance where possible.
• Test coverage should remain the same, except for added coverage in C++11 and up.

Checklists

Testing

• Bug fixes have regression tests that would reproduce the bug.
• New features should have correctness tests to validate behavior.
• Benchmarks have been added to monitor performance of new features. (not required)

[griwes]

I assume that the bit fiddling in here works and is verified by tests.

In addition to the inline comment - rename the tests in libcxx/test to match the names that actually ended up in the standard.