roundup_power_of_two performance

[SiarheiVolkau]

It seems that roundup_power_of_two is over-engineered.

Boiling down such thing leads to typical count leading zeros (clz) problem.
Modern CPUs will do it in a few cycles via special instruction(s),
we just need to use __builtin_clz for suitable compiler or rewrite it
in a way that allow compilers detect clz pattern in the code.

In current implementation no one compiler detects clz pattern thus it is just slow.

[shefty]

We need a solution that works across multiple compilers and operating systems. Is this call used anywhere in a fast path, where saving a few instructions matters?

[SiarheiVolkau]

Why don't just do like
#if defined(__GNUC__)
__builtin_clzll approach
#elif defined(ANOTHER_COMPILER_WE_WANT_TO_SUPPORT_AND_WE_KNOW_HOW)
shortcut for that compiler
#else
current implementation of roundup_power_of_two
#endif

any objections?

[shefty]

No objections