fma、fmaf、fmalfma, fmaf, fmal

9/1/2020

本文内容

将两个值相乘，添加第三个值，然后对结果进行舍入，而不会因中间舍入而丢失任何精度。Multiplies two values together, adds a third value, and then rounds the result, without losing any precision due to intermediary rounding.

语法Syntax

double fma(

double x,

double y,

double z

);

float fma(

float x,

float y,

float z

); //C++ only

long double fma(

long double x,

long double y,

long double z

); //C++ only

float fmaf(

float x,

float y,

float z

);

long double fmal(

long double x,

long double y,

long double z

);

#define fma(X, Y, Z) // Requires C11 or higher

参数Parameters

x-bladex

要相乘的第一个值。The first value to multiply.

误差y

要相乘的第二个值。The second value to multiply.

zz

要相加的值。The value to add.

返回值Return Value

返回 (x * y) + z。Returns (x * y) + z. 然后使用当前舍入格式舍入返回值。The return value is then rounded using the current rounding format.

否则，可能返回以下值之一：Otherwise, may return one of the following values:

问题Issue

返回Return

x = 无限大、 y = 0 或x = INFINITY, y = 0 or

x = 0、 y = 无限大x = 0, y = INFINITY

NaNNaN

x 或 y = 精确的±无限大， z = 无穷大，正负号x or y = exact ± INFINITY, z = INFINITY with the opposite sign

NaNNaN

x 或 y = NaNx or y = NaN

NaNNaN

not (x = 0， y= 不定) ， z = NaNnot (x = 0, y= indefinite) and z = NaN

not (x= 不定， y= 0) ， z = NaNnot (x=indefinite, y=0) and z = NaN

NaNNaN

溢出范围错误Overflow range error

± HUGE_VAL、± HUGE_VALF 或± HUGE_VALL±HUGE_VAL, ±HUGE_VALF, or ±HUGE_VALL

下溢范围错误Underflow range error

舍入后的正确值。correct value, after rounding.

按 _matherr 中所指定的报告错误。Errors are reported as specified in _matherr.

备注Remarks

由于 c + + 允许重载，因此你可以调用采用并返回和类型的 fma 的重载 float long double 。Because C++ allows overloading, you can call overloads of fma that take and return float and long double types. 在 C 程序中，除非使用 宏调用此函数，否则 fma 始终采用并返回 double 。In a C program, unless you're using the macro to call this function, fma always takes and returns a double.

如果使用 fma() 宏，则参数的类型将决定选择哪个版本的函数。If you use the fma() macro, the type of the argument determines which version of the function is selected. 有关详细信息，请参阅 类型-泛型数学 。

此函数计算值就好像它采取了无限精度，然后将最终结果舍入。This function computes the value as though it were taken to infinite precision, and then rounds the final result.

默认情况下，此函数的全局状态的作用域限定为应用程序。By default, this function's global state is scoped to the application. 若要更改此项，请参阅 CRT 中的全局状态。

要求Requirements

函数Function

C 标头C header

C++ 标头C++ header

fma、 fmaf、 fmalfma, fmaf, fmal

fma 宏fma macro

有关其他兼容性信息，请参阅兼容性。For additional compatibility information, see Compatibility.

另请参阅See also