rec

eve::rec = {}

inlineconstexpr

Computes the inverse of the parameter.

Defined in Header

#include <eve/module/core.hpp>

Callable Signatures

namespace eve
{
   template< eve::value T >
   T rec(T x) noexcept;
}

Parameters

• x : real or complex argument.

Return value

value containing the elementwise inverse value of x.

Note

For real integral value x is semantically equivalent to:

• If x==1 or x==-1, x is returned.
• If x==0, the greatest representable positive value is returned.
• Otherwise 0 is returned.

Example

#include <eve/module/core.hpp>
#include <eve/wide.hpp>
#include <iostream>
 
using wide_ft = eve::wide<float, eve::fixed<8>>;
 
int main()
{
  wide_ft pf = {0.0f, 1.0f, 2.0f, 10.0f, 200.0f, 1000.0f,
               eve::inf(eve::as<float>()), eve::nan(eve::as<float>())};
 
  std::cout << "---- simd" << '\n'
            << "<- pf            = " << pf << '\n'
            << "-> rec)(pf)      = " << eve::rec(pf) << '\n';
 
  float xf = 1.0f;
 
  std::cout << "---- scalar" << '\n'
            << "<- xf      = " << xf << '\n'
            << "-> rec(xf) = " << eve::rec(xf) << '\n';
  return 0;
}

Semantic Modifiers

• Masked Call

  The call eve;rec[mask](x) provides a masked version of eve::rec which is equivalent to if_else (mask, rec(x), x).

  Example

  #include <eve/module/core.hpp>
  #include <eve/wide.hpp>
  #include <iostream>
   
  using wide_ft = eve::wide<float, eve::fixed<8>>;
   
  int main()
  {
    wide_ft pf = {0.0f, 1.0f, 2.0f, 10.0f, 200.0f, 1000.0f,
                 eve::inf(eve::as<float>()), eve::nan(eve::as<float>())};
   
    std::cout << "---- simd" << '\n'
              << "<- pf            = " << pf << '\n'
              << "-> rec[pf > 1](pf)      = " << eve::rec[pf > 1](pf) << '\n';
    return 0;
  }

• eve::raw

  The call raw(rec)(x), call a proper system intrinsic if one exists, but with possibly very poor accuracy in return. Otherwise it uses the non decorated call.