Fixed point numbers *
Operation done when Fixed Point operations result in a value outside representable bounds.
Fixed point rounding mode.
Math operations API.
Math operations API. This object should NOT define any implementations - it should rely on BigIP classes for this. The purpose of this is to allow hardware targets to customize implementation of these operations under a common API.
Use saturating math (don't wrap).
Drop the least significant bits (always round to negative infinity).
Use a LFSR for stochastic (unbiased) rounding.
Use standard overflow wrapping on overflow or underflow.