numeric::detail::shifting_constants< FloatingType > Struct Template Reference

Helper struct with common constants needed by the floating <--> decimal conversions. More...

#include <floating_conversion.hpp>

Public Types
using	IntegerRep = std::conditional_t< is_double, uint64_t, uint32_t >
	Integer type that can hold the value of the significand.
using	ShiftingRep = std::conditional_t< is_double, __uint128_t, uint64_t >
	Shift data back and forth in space of a type with 2x the starting bits, to give us enough room.

Static Public Attributes
static constexpr bool	is_double = cuda::std::is_same_v<FloatingType, double>
	Whether the type is double.
static constexpr auto	num_significand_bits = cuda::std::numeric_limits<FloatingType>::digits
	Num bits needed to hold the significand.
static constexpr int	num_2s_shift_buffer_bits = is_double ? 4 : 1
static constexpr int	max_digits_shift = is_double ? 18 : 9
	Max at-once decimal place shift.
static constexpr int	max_bits_shift = max_digits_shift * 10 / 3
	Max at-once bit shift.
static constexpr auto	max_digits_shift_pow
	10^max_digits_shift More...

Detailed Description

template<typename FloatingType>
struct numeric::detail::shifting_constants< FloatingType >

Helper struct with common constants needed by the floating <--> decimal conversions.

Definition at line 557 of file floating_conversion.hpp.

Member Data Documentation

max_digits_shift_pow

template<typename FloatingType >

constexpr auto numeric::detail::shifting_constants< FloatingType >::max_digits_shift_pow

staticconstexpr

Initial value:

=
    multiply_power10<IntegerRep>(IntegerRep(1), max_digits_shift)

10^max_digits_shift

Definition at line 618 of file floating_conversion.hpp.

num_2s_shift_buffer_bits

template<typename FloatingType >

constexpr int numeric::detail::shifting_constants< FloatingType >::num_2s_shift_buffer_bits = is_double ? 4 : 1

staticconstexpr

buffer bits to account for shifting error

Definition at line 601 of file floating_conversion.hpp.

---

The documentation for this struct was generated from the following file:

• floating_conversion.hpp