This PR fixes the implementation of intrinsic_tsc on AARCH64 to match the scale of values returned on X86_64 and switch to leverage special registries to fetch the value instead of trying to calculate them.
Also rehaul the fallback mechanism to calculate the cpu clocks per second to make it more reliable and precise (and faster).
This PR fixes the implementation of intrinsic_tsc on AARCH64 to match the scale of values returned on X86_64 and switch to leverage special registries to fetch the value instead of trying to calculate them.
Also rehaul the fallback mechanism to calculate the cpu clocks per second to make it more reliable and precise (and faster).