_mm256_maddubs_epi16
Classification
AVX_ALL, Arithmetic, CPUID Test: AVX2
Header File
immintrin.h
Instruction
VPMADDUBSW ymm, ymm, ymm
Synopsis
 _mm256_maddubs_epi16(__m256i a, __m256i b);
Description
Vertically multiply each unsigned 8-bit integer from "a" with the corresponding signed 8-bit integer from "b", producing intermediate signed 16-bit integers. Horizontally add adjacent pairs of intermediate signed 16-bit integers, and pack the saturated results in "dst".
Operation
FOR j := 0 to 15
	i := j*16
	dst[i+15:i] := Saturate16( a[i+15:i+8]*b[i+15:i+8] + a[i+7:i]*b[i+7:i] )
ENDFOR
dst[MAX:256] := 0