significants

eve::significants = {}

inlineconstexpr

Computes the rounding to n significants digits of the first input.

Defined in Header

#include <eve/module/core.hpp>

Callable Signatures

namespace eve
{
   template< eve::value T, eve::integral_value N >
   eve::common_compatible_t<T, N> significants(T x, N n) noexcept;
}

Parameters

• x: floating argument.
• n : integral value argument.

Return value

Computes elementwise the rounding to n significants digits of x. With null n the result is a NaN.

Warning

Floating numbers are not stored in decimal form. So if you try significants with a not exactly representable number the result can be not exactly what you expect.

Example

#include <eve/module/math.hpp>
#include <eve/wide.hpp>
#include <iostream>
#include <iomanip>
 
using iT      = std::int32_t;
using wide_it = eve::wide<iT, eve::fixed<4>>;
using wide_ft = eve::wide<double, eve::fixed<4>>;
 
int main()
{
  wide_it qi = {0, 1, 2, 4};
  wide_ft pf = {1.2345678901f, 1.2345678901f, 1.2345678901f, 1.2345678901f};
 
  std::cout << "---- simd" << std::setprecision(8) << '\n'
            << "<- pf            = " << pf << '\n'
            << "<- qi            = " << qi << '\n'
            << "-> significants(pf, qi) = " << eve::significants(pf, qi) << '\n';
 
  float xf = 2.34316;
  iT yi = 3;
 
  std::cout << "---- scalar" << '\n'
            << "<- xf             = " << xf << '\n'
            << "<- yi             = " << yi << '\n'
            << "-> significants(xf, yi)  = " << eve::significants(xf, yi) << '\n';
  return 0;
}

Semantic Modifiers

• Masked Call

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

  Example