Added more vector intrinsics for AVX512 and support for left-packing vectors in multiple instruction sets.