Lines Matching refs:BPS
407 vst1_lane_u32((uint32_t*)(dst + 0 * BPS), vreinterpret_u32_u8(dst01_u8), 0); in SaturateAndStore4x4()
408 vst1_lane_u32((uint32_t*)(dst + 1 * BPS), vreinterpret_u32_u8(dst01_u8), 1); in SaturateAndStore4x4()
409 vst1_lane_u32((uint32_t*)(dst + 2 * BPS), vreinterpret_u32_u8(dst23_u8), 0); in SaturateAndStore4x4()
410 vst1_lane_u32((uint32_t*)(dst + 3 * BPS), vreinterpret_u32_u8(dst23_u8), 1); in SaturateAndStore4x4()
419 dst01 = vld1_lane_u32((uint32_t*)(dst + 0 * BPS), dst01, 0); in Add4x4()
420 dst23 = vld1_lane_u32((uint32_t*)(dst + 2 * BPS), dst23, 0); in Add4x4()
421 dst01 = vld1_lane_u32((uint32_t*)(dst + 1 * BPS), dst01, 1); in Add4x4()
422 dst23 = vld1_lane_u32((uint32_t*)(dst + 3 * BPS), dst23, 1); in Add4x4()
1041 const int kBPS = BPS; in TransformOne()
1270 const uint8x8_t A = vld1_u8(dst - BPS); // top row in DC4()
1273 const uint16x8_t L0 = vmovl_u8(vld1_u8(dst + 0 * BPS - 1)); in DC4()
1274 const uint16x8_t L1 = vmovl_u8(vld1_u8(dst + 1 * BPS - 1)); in DC4()
1275 const uint16x8_t L2 = vmovl_u8(vld1_u8(dst + 2 * BPS - 1)); in DC4()
1276 const uint16x8_t L3 = vmovl_u8(vld1_u8(dst + 3 * BPS - 1)); in DC4()
1285 vst1_lane_u32((uint32_t*)(dst + i * BPS), vreinterpret_u32_u8(dc), 0); in DC4()
1291 const uint8x8_t TL = vld1_dup_u8(dst - BPS - 1); // top-left pixel 'A[-1]' in TrueMotion()
1292 const uint8x8_t T = vld1_u8(dst - BPS); // top row 'A[0..3]' in TrueMotion()
1297 const int16x8_t L0 = ConvertU8ToS16(vld1_dup_u8(dst + 0 * BPS - 1)); in TrueMotion()
1298 const int16x8_t L1 = ConvertU8ToS16(vld1_dup_u8(dst + 1 * BPS - 1)); in TrueMotion()
1299 const int16x8_t L2 = ConvertU8ToS16(vld1_dup_u8(dst + 2 * BPS - 1)); in TrueMotion()
1300 const int16x8_t L3 = ConvertU8ToS16(vld1_dup_u8(dst + 3 * BPS - 1)); in TrueMotion()
1311 vst1_lane_u32((uint32_t*)(dst + 0 * BPS), r0_u32, 0); in TrueMotion()
1312 vst1_lane_u32((uint32_t*)(dst + 1 * BPS), r1_u32, 0); in TrueMotion()
1313 vst1_lane_u32((uint32_t*)(dst + 2 * BPS), r2_u32, 0); in TrueMotion()
1314 vst1_lane_u32((uint32_t*)(dst + 3 * BPS), r3_u32, 0); in TrueMotion()
1316 vst1_u32((uint32_t*)(dst + 0 * BPS), r0_u32); in TrueMotion()
1317 vst1_u32((uint32_t*)(dst + 1 * BPS), r1_u32); in TrueMotion()
1318 vst1_u32((uint32_t*)(dst + 2 * BPS), r2_u32); in TrueMotion()
1319 vst1_u32((uint32_t*)(dst + 3 * BPS), r3_u32); in TrueMotion()
1321 dst += 4 * BPS; in TrueMotion()
1329 const uint64x1_t A0 = vreinterpret_u64_u8(vld1_u8(dst - BPS - 1)); // top row in VE4()
1339 vst1_lane_u32((uint32_t*)(dst + i * BPS), vreinterpret_u32_u8(avg), 0); in VE4()
1344 const uint8x8_t XABCD_u8 = vld1_u8(dst - BPS - 1); in RD4()
1347 const uint32_t I = dst[-1 + 0 * BPS]; in RD4()
1348 const uint32_t J = dst[-1 + 1 * BPS]; in RD4()
1349 const uint32_t K = dst[-1 + 2 * BPS]; in RD4()
1350 const uint32_t L = dst[-1 + 3 * BPS]; in RD4()
1365 vst1_lane_u32((uint32_t*)(dst + 0 * BPS), r0, 0); in RD4()
1366 vst1_lane_u32((uint32_t*)(dst + 1 * BPS), r1, 0); in RD4()
1367 vst1_lane_u32((uint32_t*)(dst + 2 * BPS), r2, 0); in RD4()
1368 vst1_lane_u32((uint32_t*)(dst + 3 * BPS), r3, 0); in RD4()
1373 const uint8x8_t ABCDEFGH = vld1_u8(dst - BPS + 0); in LD4()
1374 const uint8x8_t BCDEFGH0 = vld1_u8(dst - BPS + 1); in LD4()
1375 const uint8x8_t CDEFGH00 = vld1_u8(dst - BPS + 2); in LD4()
1376 const uint8x8_t CDEFGHH0 = vset_lane_u8(dst[-BPS + 7], CDEFGH00, 6); in LD4()
1384 vst1_lane_u32((uint32_t*)(dst + 0 * BPS), r0, 0); in LD4()
1385 vst1_lane_u32((uint32_t*)(dst + 1 * BPS), r1, 0); in LD4()
1386 vst1_lane_u32((uint32_t*)(dst + 2 * BPS), r2, 0); in LD4()
1387 vst1_lane_u32((uint32_t*)(dst + 3 * BPS), r3, 0); in LD4()
1394 const uint8x8_t top = vld1_u8(dst - BPS); in VE8uv()
1397 vst1_u8(dst + j * BPS, top); in VE8uv()
1406 dst += BPS; in HE8uv()
1416 const uint8x8_t A = vld1_u8(dst - BPS); // top row in DC8()
1424 const uint16x8_t L0 = vmovl_u8(vld1_u8(dst + 0 * BPS - 1)); in DC8()
1425 const uint16x8_t L1 = vmovl_u8(vld1_u8(dst + 1 * BPS - 1)); in DC8()
1426 const uint16x8_t L2 = vmovl_u8(vld1_u8(dst + 2 * BPS - 1)); in DC8()
1427 const uint16x8_t L3 = vmovl_u8(vld1_u8(dst + 3 * BPS - 1)); in DC8()
1428 const uint16x8_t L4 = vmovl_u8(vld1_u8(dst + 4 * BPS - 1)); in DC8()
1429 const uint16x8_t L5 = vmovl_u8(vld1_u8(dst + 5 * BPS - 1)); in DC8()
1430 const uint16x8_t L6 = vmovl_u8(vld1_u8(dst + 6 * BPS - 1)); in DC8()
1431 const uint16x8_t L7 = vmovl_u8(vld1_u8(dst + 7 * BPS - 1)); in DC8()
1456 vst1_u32((uint32_t*)(dst + i * BPS), vreinterpret_u32_u8(dc)); in DC8()
1472 const uint8x16_t top = vld1q_u8(dst - BPS); in VE16()
1475 vst1q_u8(dst + j * BPS, top); in VE16()
1484 dst += BPS; in HE16()
1494 const uint8x16_t A = vld1q_u8(dst - BPS); // top row in DC16()
1506 const uint16x8_t L0 = vmovl_u8(vld1_u8(dst + (i + 0) * BPS - 1)); in DC16()
1507 const uint16x8_t L1 = vmovl_u8(vld1_u8(dst + (i + 1) * BPS - 1)); in DC16()
1508 const uint16x8_t L2 = vmovl_u8(vld1_u8(dst + (i + 2) * BPS - 1)); in DC16()
1509 const uint16x8_t L3 = vmovl_u8(vld1_u8(dst + (i + 3) * BPS - 1)); in DC16()
1510 const uint16x8_t L4 = vmovl_u8(vld1_u8(dst + (i + 4) * BPS - 1)); in DC16()
1511 const uint16x8_t L5 = vmovl_u8(vld1_u8(dst + (i + 5) * BPS - 1)); in DC16()
1512 const uint16x8_t L6 = vmovl_u8(vld1_u8(dst + (i + 6) * BPS - 1)); in DC16()
1513 const uint16x8_t L7 = vmovl_u8(vld1_u8(dst + (i + 7) * BPS - 1)); in DC16()
1540 vst1q_u8(dst + i * BPS, dc); in DC16()
1551 const uint8x8_t TL = vld1_dup_u8(dst - BPS - 1); // top-left pixel 'A[-1]' in TM16()
1552 const uint8x16_t T = vld1q_u8(dst - BPS); // top row 'A[0..15]' in TM16()
1559 const int16x8_t L0 = ConvertU8ToS16(vld1_dup_u8(dst + 0 * BPS - 1)); in TM16()
1560 const int16x8_t L1 = ConvertU8ToS16(vld1_dup_u8(dst + 1 * BPS - 1)); in TM16()
1561 const int16x8_t L2 = ConvertU8ToS16(vld1_dup_u8(dst + 2 * BPS - 1)); in TM16()
1562 const int16x8_t L3 = ConvertU8ToS16(vld1_dup_u8(dst + 3 * BPS - 1)); in TM16()
1576 vst1q_u8(dst + 0 * BPS, row0); in TM16()
1577 vst1q_u8(dst + 1 * BPS, row1); in TM16()
1578 vst1q_u8(dst + 2 * BPS, row2); in TM16()
1579 vst1q_u8(dst + 3 * BPS, row3); in TM16()
1580 dst += 4 * BPS; in TM16()