PMULHW - Packed MULtiply High Word
PMULHW xmm1, xmm2/m128 (S2
__m128i _mm_mulhi_epi16(__m128i a, __m128i b)
For each WORD, calculate signed (1) * signed (2) and set the upper 16-bit of the 32-bit result to (3).
VPMULHW xmm1, xmm2, xmm3/m128 (V1
__m128i _mm_mulhi_epi16(__m128i a, __m128i b)
VPMULHW xmm1{k1}{z}, xmm2, xmm3/m128 (V5+BW+VL
__m128i _mm_mask_mulhi_epi16(__m128i s, __mmask8 k, __m128i a, __m128i b)
__m128i _mm_maskz_mulhi_epi16(__mmask8 k, __m128i a, __m128i b)
For each WORD, calculate signed (1) * signed (2) and set the upper 16-bit of the 32-bit result to (3).
VPMULHW ymm1, ymm2, ymm3/m256 (V2
__m256i _mm256_mulhi_epi16(__m256i a, __m256i b)
VPMULHW ymm1{k1}{z}, ymm2, ymm3/m256 (V5+BW+VL
__m256i _mm256_mask_mulhi_epi16(__m256i s, __mmask16 k, __m256i a, __m256i b)
__m256i _mm256_maskz_mulhi_epi16(__mmask16 k, __m256i a, __m256i b)
For each WORD, calculate signed (1) * signed (2) and set the upper 16-bit of the 32-bit result to (3).
VPMULHW zmm1{k1}{z}, zmm2, zmm3/m512 (V5+BW
__m512i _mm512_mulhi_epi16(__m512i a, __m512i b)
__m512i _mm512_mask_mulhi_epi16(__m512i s, __mmask32 k, __m512i a, __m512i b)
__m512i _mm512_maskz_mulhi_epi16(__mmask32 k, __m512i a, __m512i b)
For each WORD, calculate signed (1) * signed (2) and set the upper 16-bit of the 32-bit result to (3).
x86/x64 SIMD Instruction List
Feedback