Decimal WAD

Math for preserving precision floats which are limited to be at most u64::MAX.