#define vec_add_16(a,b) _mm512_add_epi16(a,b)
#define vec_sub_16(a,b) _mm512_sub_epi16(a,b)
static constexpr IndexType kNumRegs = 8; // only 8 are needed
#elif USE_AVX2
typedef __m256i vec_t;
#define vec_add_16(a,b) _mm512_add_epi16(a,b)
#define vec_sub_16(a,b) _mm512_sub_epi16(a,b)
static constexpr IndexType kNumRegs = 8; // only 8 are needed
#elif USE_AVX2
typedef __m256i vec_t;