Sentinel floating-point conversion functions

Microsoft Specific

Conversion functions between floating-point types and integral types that use an Intel Architecture (IA) AVX-512 compatible sentinel strategy.

Syntax

int _cvt_ftoi_sent(float value);
long long _cvt_ftoll_sent(float value);
unsigned _cvt_ftoui_sent(float value);
unsigned long long _cvt_ftoull_sent(float value);
int _cvt_dtoi_sent(double value);
long long _cvt_dtoll_sent(double value);
unsigned _cvt_dtoui_sent(double value);
unsigned long long _cvt_dtoull_sent(double value);

Parameters

value
[in] A floating-point value to convert.

Return value

The integer-typed result of the conversion.

Requirements

Header: <intrin.h>

Architecture: x86, x64

Remarks

These intrinsics are floating-point to integral type conversion functions that use a sentinel strategy: They return the result value farthest from zero as a proxy sentinel value for NaN. Any invalid conversion returns this sentinel value. The specific sentinel value returned depends on the result type.

The sentinel conversion intrinsics are available starting in Visual Studio 2019 version 16.10.

END Microsoft Specific

See also

Compiler intrinsics
Fast floating-point conversion functions
Saturation floating-point conversion functions