fma

Multiply and add using fused multiply add approach

Syntax

X = fma(A, B, C)

Description

X = fma(A, B, C) computes A.*B+C using a fused multiply add approach. Fused multiply add operations round only once, often making the result more accurate than performing a multiplication operation followed by an addition.

Examples

Multiply and Add Three Inputs Using Fused Multiply Add

This example shows how to use the fma function to calculate $$A\times B+C$$ using a fused multiply add approach.

Define the inputs and use the fma function to compute the multiply add operation.

a = half(10);
b = half(10);
c = half(2);
x = fma(a, b, c)

x = 

  half

   102

Compare the result of the fma function with the two-step approach of computing the product and then the sum.

temp = a * b;
x = temp + c

x = 

  half

   102

Input Arguments

A — Input array

scalar | vector | matrix | multidimensional array

Input array, specified as a floating-point scalar, vector, matrix, or multidimensional array. When A and B are matrices, fma performs element-wise multiplication followed by addition.

Data Types: single | double | half

B — Input array

scalar | vector | matrix | multidimensional array

Input array, specified as a floating-point scalar, vector, matrix, or multidimensional array. When A and B are matrices, fma performs element-wise multiplication followed by addition.

Data Types: single | double | half

C — Input array

scalar | vector | matrix | multidimensional array

Input array, specified as a floating-point scalar, vector, matrix, or multidimensional array.

Data Types: single | double | half

Output Arguments

X — Result of multiply and add operation

scalar | vector | matrix | multidimensional array

Result of multiply and add operation, A.*B+C, returned as a scalar, vector, matrix, or multidimensional array.

Version History

Introduced in R2019a