Intel® C++ Compiler 16.0 User and Reference Guide

_mm512_mask_blend_ps

Blends float32 vectors using the instruction mask. Corresponding instruction is VBLENDMPS. This intrinsic only applies to Intel® Many Integrated Core Architecture (Intel® MIC Architecture).

Syntax

extern __m512 __cdecl _mm512_mask_blend_ps(__mmask16 k1, __m512 v2, __m512 v3);

Parameters

v2

float32 vector used for the blend operation

v3

float32 vector used for the blend operation

k1

instruction mask used for blend selection

Description

This intrinsic performs an element-by-element blending between float32 vectors v2 and v3, using mask k1 as selector. The result is written to float32 vector v2.

The k1 mask is not used as a write-mask for this instruction. Instead, the k1 mask is used as an element selector: every element of the destination is conditionally selected between vectors v2 or v3 using the value of the related mask bit ('0' for source vector v2, '1' for source vector v3).

Returns

Returns the result of the blend operation.