simdpp::rcp_rh

From libsimdpp-docs
< fp
template<unsigned N, class E>
Ret<N, _DETAIL_> rcp_rh( const Vec<N,E>& e, const Vec<N,E>& x );

Performs single Newton-Rhapson iteration improving the current estimate e of reciprocal of x. Effectively implements e * (2 - e*x).

The implementation behaves as if the following set of overloads is provided:

Ret Vec
float32 float32
float64 float64

Parameters[edit]

e - current estimate
x - values to compute reciprocal for

Return value[edit]

A vector expression evaluating to e * (2 - e*x).

See also[edit]