Performance improvements

[danakj]

Use google benchmark or other performance/profiling tools to make better choices about inlining, branch ordering, [[likely]]/[[unlikely]] or other code gen in critical paths to improve performance.

[danakj]

Use clang's optimization reports to find opportunities: https://johnnysswlab.com/loop-optimizations-interpreting-the-compiler-optimization-report/

[danakj]

In chrome every time we un-inline something performance goes up? Or frequently? So probably a lot of attribute((noinline)) is in order, driven by some kinda measurements.

That said libc++ is firmly against noinline, and they only use out-of-line/out-of-header non-template implementations to move code away from being inlined.

[danakj]

Benchmarking framework added in #305.

Benchmarks go in the bench/ directory and use GoogleTest for now to register each function to run.

They are not part of the tests discovered by CMake, and run by ctest, you need to run out/bench/bench directly to run them.