Developer Guide and Reference

Contents

_mm256_shufflelo_epi16

Shuffles the low 4 signed or unsigned words in each 128-bit lane of the source operand according to the shuffle control operand. The high qwords in each of 2 128-bit lanes of the source operand are copied to the corresponding high qwords of the result value. The corresponding Intel® AVX 2 instruction is
VPSHUFLW
.

Syntax

extern __m256i _mm256_shufflelo_epi16(__m256i val, const int control);
Arguments
val
integer source vector
control
immediate control operand
Description
Shuffles the low four signed or unsigned words in each 128-bit lane of the source operand according to the shuffle control operand. The high qwords in each of 2 128-bit lanes of the source operand are copied to the corresponding high qwords of the result value. The shuffle value must be an immediate.
Below is the pseudo-code for the intrinsic:
RESULT[15:0] <- (VAL >> (CONTROL[1:0] *16))[15:0] RESULT[31:16] <- (VAL >> (CONTROL[3:2] * 16))[15:0] RESULT[47:32] <- (VAL >> (CONTROL[5:4] * 16))[15:0] RESULT[63:48] <- (VAL >> (CONTROL[7:6] * 16))[15:0] RESULT[127:64] <- VAL[127:64] RESULT[143:128] <- (VAL >> (CONTROL[1:0] *16))[143:128] RESULT[159:144] <- (VAL >> (CONTROL[3:2] * 16))[143:128] RESULT[175:160] <- (VAL >> (CONTROL[5:4] * 16))[143:128] RESULT[191:176] <- (VAL >> (CONTROL[7:6] * 16))[143:128] RESULT[255:192] <- VAL[255:192]
Returns
Result of the shuffle operation.