54   vst1_lane_u32((uint32_t*)(dst + 0 * BPS), vreinterpret_u32_u8(dst01_u8), 0);  in SaturateAndStore4x4()
55   vst1_lane_u32((uint32_t*)(dst + 1 * BPS), vreinterpret_u32_u8(dst01_u8), 1);  in SaturateAndStore4x4()
56   vst1_lane_u32((uint32_t*)(dst + 2 * BPS), vreinterpret_u32_u8(dst23_u8), 0);  in SaturateAndStore4x4()
57   vst1_lane_u32((uint32_t*)(dst + 3 * BPS), vreinterpret_u32_u8(dst23_u8), 1);  in SaturateAndStore4x4()
66   dst01 = vld1_lane_u32((uint32_t*)(ref + 0 * BPS), dst01, 0);  in Add4x4()
67   dst23 = vld1_lane_u32((uint32_t*)(ref + 2 * BPS), dst23, 0);  in Add4x4()
68   dst01 = vld1_lane_u32((uint32_t*)(ref + 1 * BPS), dst01, 1);  in Add4x4()
69   dst23 = vld1_lane_u32((uint32_t*)(ref + 3 * BPS), dst23, 1);  in Add4x4()
132   const int kBPS = BPS;  in ITransformOne()
257   out = vld1q_lane_u32((const uint32_t*)(src + 0 * BPS), out, 0);  in Load4x4()
258   out = vld1q_lane_u32((const uint32_t*)(src + 1 * BPS), out, 1);  in Load4x4()
259   out = vld1q_lane_u32((const uint32_t*)(src + 2 * BPS), out, 2);  in Load4x4()
260   out = vld1q_lane_u32((const uint32_t*)(src + 3 * BPS), out, 3);  in Load4x4()
363   const int kBPS = BPS;  in FTransform()
638   LOAD_LANE_32b(a + 0 * BPS, d0d1, 0);  // a00 a01 a02 a03  in Disto4x4()
639   LOAD_LANE_32b(a + 1 * BPS, d0d1, 1);  // a10 a11 a12 a13  in Disto4x4()
640   LOAD_LANE_32b(b + 0 * BPS, d0d1, 2);  // b00 b01 b02 b03  in Disto4x4()
641   LOAD_LANE_32b(b + 1 * BPS, d0d1, 3);  // b10 b11 b12 b13  in Disto4x4()
642   LOAD_LANE_32b(a + 2 * BPS, d2d3, 0);  // a20 a21 a22 a23  in Disto4x4()
643   LOAD_LANE_32b(a + 3 * BPS, d2d3, 1);  // a30 a31 a32 a33  in Disto4x4()
644   LOAD_LANE_32b(b + 2 * BPS, d2d3, 2);  // b20 b21 b22 b23  in Disto4x4()
645   LOAD_LANE_32b(b + 3 * BPS, d2d3, 3);  // b30 b31 b32 b33  in Disto4x4()
683   const int kBPS = BPS;  in Disto4x4()
873   for (y = 0; y < 16 * BPS; y += 4 * BPS) {  in Disto16x16()
935     AccumulateSSE16(a + y * BPS, b + y * BPS, &sum);  in SSE16x16()
944     AccumulateSSE16(a + y * BPS, b + y * BPS, &sum);  in SSE16x8()
953     const uint8x8_t a0 = vld1_u8(a + y * BPS);  in SSE8x8()
954     const uint8x8_t b0 = vld1_u8(b + y * BPS);  in SSE8x8()