Intrinsics for Pack/Unpack Operations
- _mm256_packs_epi16/32
Pack signed word/doubleword integers to signed byte/words integers and saturates. The corresponding Intel® AVX2 instruction is VPACKSSWB or VPACKSSDW.
- _mm256_packus_epi16/32
Pack signed word/doubleword integers to unsigned byte/word integers and saturates. The corresponding Intel® AVX2 instruction is VPACKUSWB or VPACKUSDW.
- _mm256_unpackhi_epi8/16/32/64
Unpacks and interleaves the high-order data elements of the source vector with the high-order data elements in the destination vector. The corresponding Intel® AVX2 instruction is VPUNPCKHBW, VPUNPCKHWD, VPUNPCKHDQ, or VPUNPCKHQDQ.
- _mm256_unpacklo_epi8/16/32/64
Unpacks and interleaves the low-order data elements of the source vector with the low-order data elements in the destination vector. The corresponding Intel® AVX2 instruction is VPUNPCKLBW, VPUNPCKLWD, VPUNPCKLDQ, or VPUNPCKLQDQ.