std::numeric_limits<T>::round_style

From cppreference.com
 
C++
 
Utilities library
Type support (basic types, RTTI, type traits)
Dynamic memory management
Error handling
Program utilities
Variadic functions
Library feature-test macros
Date and time
Function objects
(C++11)
(C++11)
(C++14)
Relational operators (deprecated in C++20)
Comparisons (C++20)
Common vocabulary types
(C++11)
(C++17)
(C++17)
(C++17)
(C++17)
(C++17)

Swap, forward and move
(C++14)
(C++11)
(C++11)
(C++11)
Elementary string conversions
(C++17)
(C++17)
(C++17)
Type operations
(C++11)
(C++17)
(C++17)
 
Type support
Basic types
Fundamental types
Fixed width integer types (C++11)
Numeric limits
C numeric limits interface
Runtime type information
Type traits
Type categories
(C++11)
(C++14)
(C++11)
(C++11)
(C++11)
(C++11)
(C++11)
(C++11)
(C++11)
(C++11)
(C++11)
(C++11)
(C++11)
(C++11)
(C++11)
(C++11)
(C++11)
(C++11)
(C++11)
(C++11)
(C++11)
Type properties
(C++11)
(C++11)
(C++11)
(C++11)
(C++14)
(C++11)
(C++17)
(C++11)
(C++11)
(C++11)
(C++11)(until C++20)
(C++11)(deprecated in C++20)
(C++11)
(C++11)
(C++20)
(C++20)
(C++17)
Type trait constants
(C++11)(C++17)(C++11)(C++11)
Metafunctions
(C++17)
(C++17)
(C++17)
Endian
(C++20)
Constant evaluation context
(C++20)
Supported operations
(C++11)(C++11)(C++11)
(C++11)(C++11)(C++11)
(C++11)(C++11)(C++11)
(C++11)(C++11)(C++11)
(C++11)(C++11)(C++11)
(C++11)(C++11)(C++11)
(C++11)(C++11)(C++11)
(C++11)(C++11)(C++11)
(C++11)
(C++17)(C++17)(C++17)(C++17)
Relationships and property queries
(C++11)
(C++11)
(C++11)
(C++17)(C++17)(C++17)(C++17)
Type modifications
(C++11)(C++11)(C++11)
(C++11)(C++11)(C++11)
(C++11)
(C++11)
Type transformations
(C++11)
(C++11)
(C++11)
(C++20)
(C++11)
(C++17)
(C++11)
(C++11)
(C++20)
(C++11)
(C++11)(until C++20)(C++17)
(C++20)
 
std::numeric_limits
Static constants
numeric_limits::round_style
(C++11)
Static member functions
(C++11)
Helper types
 
static const std::float_round_style round_style;
 (until C++11)
static constexpr std::float_round_style round_style;
 (since C++11)

The value of std::numeric_limits<T>::round_style identifies the rounding style used by the floating-point type T whenever a value that is not one of the exactly repesentable values of T is stored in an object of that type.

Standard specializations

T value of std::numeric_limits<T>::round_style
/* non-specialized */ std::round_toward_zero
bool std::round_toward_zero
char std::round_toward_zero
signed char std::round_toward_zero
unsigned char std::round_toward_zero
wchar_t std::round_toward_zero
char8_t std::round_toward_zero
char16_t std::round_toward_zero
char32_t std::round_toward_zero
short std::round_toward_zero
unsigned short std::round_toward_zero
int std::round_toward_zero
unsigned int std::round_toward_zero
long std::round_toward_zero
unsigned long std::round_toward_zero
long long std::round_toward_zero
unsigned long long std::round_toward_zero
float usually std::round_to_nearest
double usually std::round_to_nearest
long double usually std::round_to_nearest

Notes

These values are constants, and do not reflect the changes to the rounding made by std::fesetround. The changed values may be obtained from FLT_ROUNDS or std::fegetround.

Example

The decimal value 0.1 cannot be represented by a binary floating-point type. When stored in an IEEE-745 double, it falls between 0x1.9999999999999*2-4
 and 0x1.999999999999a*2-4
. Rounding to nearest representable value results in 0x1.999999999999a*2-4
.

Similarly, the decimal value 0.3, which is between 0x1.3333333333333*2-2
 and 0x1.3333333333334*2-2
 is rounded to nearest and is stored as 0x1.3333333333333*2-2
.

Run this code
#include <iostream>
#include <limits>
int main()
{
    std::cout << std::hexfloat << "The decimal 0.1 is stored in a double as "
              << 0.1 << '\n'
              << "The decimal 0.3 is stored in a double as "
              << 0.3 << '\n'
              << "The rounding style is " << std::numeric_limits<double>::round_style << '\n';
}

Output: 

The decimal 0.1 is stored in a double as 0x1.999999999999ap-4
The decimal 0.3 is stored in a double as 0x1.3333333333333p-2
The rounding style is 1

See also

 indicates floating-point rounding modes
(enum)
Retrieved from "https://en.cppreference.com/mwiki/index.php?title=cpp/types/numeric_limits/round_style&oldid=108352"