vdupq_n_u32
rk32 = vdupq_n_u32(vgetq_lane_u32(rk32, 3));
vdupq_n_u32(vgetq_lane_u32(prkhi32, 0)),
vdupq_n_u32(vgetq_lane_u32(prk32, 3)),
vdupq_n_u32(
vdupq_n_u32(
const uint32x4_t carry = vsetq_lane_u32(0x87, vdupq_n_u32(1), 0);
const uint32x4_t ctr32_inc = vsetq_lane_u32(1, vdupq_n_u32(0), 3);
const uint32x4_t ctr32_inc = vsetq_lane_u32(1, vdupq_n_u32(0), 3);
vsetq_lane_u32(1, vdupq_n_u32(0), 0);
vsetq_lane_u32(1, vdupq_n_u32(0), 0);