/external/libgav1/libgav1/src/dsp/x86/ |
D | loop_restoration_10bit_sse4.cc | 1438 __m128i b3[2][6], __m128i* const ma5, __m128i b5[2]) { in BoxFilterPreProcessLo() argument 1462 CalculateIntermediate5<0>(s5, sq5, scales[0], ma5, b5); in BoxFilterPreProcessLo() 1470 __m128i b3[2][6], __m128i ma5[2], __m128i b5[6]) { in BoxFilterPreProcess() 1497 CalculateIntermediate5<8>(s5[0], sq5, scales[0], &ma5[0], b5 + 2); in BoxFilterPreProcess() 1516 CalculateIntermediate5<0>(s5[1], sq5, scales[0], &ma5[1], b5 + 4); in BoxFilterPreProcess() 1523 __m128i* const ma5, __m128i b3[2], __m128i b5[2]) { in BoxFilterPreProcessLastRowLo() argument 1533 CalculateIntermediate5<0>(s5, sq5, scales[0], ma5, b5); in BoxFilterPreProcessLastRowLo() 1544 __m128i ma5[2], __m128i b3[6], __m128i b5[6]) { in BoxFilterPreProcessLastRow() 1554 CalculateIntermediate5<8>(s5[0], sq5, scales[0], ma5, b5 + 2); in BoxFilterPreProcessLastRow() 1566 CalculateIntermediate5<0>(s5[1], sq5, scales[0], ma5 + 1, b5 + 4); in BoxFilterPreProcessLastRow() [all …]
|
D | loop_restoration_sse4.cc | 1572 __m128i b3[2][3], __m128i* const ma5, __m128i* const b5) { in BoxFilterPreProcessLo() argument 1596 CalculateIntermediate5<0>(s5, sq5, scales[0], ma5, b5); in BoxFilterPreProcessLo() 1604 __m128i b3[2][3], __m128i ma5[2], __m128i b5[3]) { in BoxFilterPreProcess() 1631 CalculateIntermediate5<8>(s5[0], sq5, scales[0], &ma5[0], &b5[1]); in BoxFilterPreProcess() 1650 CalculateIntermediate5<0>(s5[1], sq5, scales[0], &ma5[1], &b5[2]); in BoxFilterPreProcess() 1657 __m128i* const ma5, __m128i* const b3, __m128i* const b5) { in BoxFilterPreProcessLastRowLo() argument 1667 CalculateIntermediate5<0>(s5, sq5, scales[0], ma5, b5); in BoxFilterPreProcessLastRowLo() 1678 __m128i ma5[2], __m128i b3[3], __m128i b5[3]) { in BoxFilterPreProcessLastRow() 1688 CalculateIntermediate5<8>(s5[0], sq5, scales[0], ma5, b5 + 1); in BoxFilterPreProcessLastRow() 1700 CalculateIntermediate5<0>(s5[1], sq5, scales[0], ma5 + 1, b5 + 2); in BoxFilterPreProcessLastRow() [all …]
|
D | loop_restoration_10bit_avx2.cc | 1998 __m128i b3[2][10], __m128i* const ma5, __m128i b5[2]) { in BoxFilterPreProcessLo() argument 2022 CalculateIntermediate5(s5, sq5, scales[0], ma5, b5); in BoxFilterPreProcessLo() 2031 __m256i b3[2][7], __m256i ma5[3], __m256i b5[5]) { in BoxFilterPreProcess() 2095 CalculateIntermediate<25>(sum_5, index_5, ma5, t, t + 2); in BoxFilterPreProcess() 2103 __m128i* const ma5, __m128i b3[2], __m128i b5[2]) { in BoxFilterPreProcessLastRowLo() argument 2113 CalculateIntermediate5(s5, sq5, scales[0], ma5, b5); in BoxFilterPreProcessLastRowLo() 2124 __m256i sq[6], __m256i ma3[2], __m256i ma5[2], __m256i b3[5], in BoxFilterPreProcessLastRow() 2161 CalculateIntermediate<25>(sum_5, index_5, ma5, t, t + 2); in BoxFilterPreProcessLastRow() 2193 __m256i ma5[3], ma[2], b[4]; in BoxSumFilterPreProcess5() local 2198 Prepare3_8(mas, ma5); in BoxSumFilterPreProcess5() [all …]
|
D | loop_restoration_avx2.cc | 1897 __m128i b3[2], __m128i* const ma5, __m128i* const b5) { in BoxFilterPreProcessLo() argument 1923 CalculateIntermediate5(s5, sq5, scales[0], ma5, b5); in BoxFilterPreProcessLo() 1932 __m256i b3[2][5], __m256i ma5[3], __m256i b5[5]) { in BoxFilterPreProcess() 1986 CalculateIntermediate<25>(sum_5, index_5, ma5, b5 + 1); in BoxFilterPreProcess() 1996 __m128i* const ma5, __m128i* const b3, __m128i* const b5) { in BoxFilterPreProcessLastRowLo() argument 2006 CalculateIntermediate5(s5, sq5, scales[0], ma5, b5); in BoxFilterPreProcessLastRowLo() 2017 __m256i sq[6], __m256i ma3[2], __m256i ma5[2], __m256i b3[5], in BoxFilterPreProcessLastRow() 2049 CalculateIntermediate<25>(sum_5, index_5, ma5, b5 + 1); in BoxFilterPreProcessLastRow() 2075 __m256i ma5[3], ma[2], b[4]; in BoxSumFilterPreProcess5() local 2079 Prepare3_8(mas, ma5); in BoxSumFilterPreProcess5() [all …]
|
/external/libgav1/libgav1/src/dsp/arm/ |
D | loop_restoration_neon.cc | 882 inline uint16x8_t Sum565(const uint8x16_t ma5[2]) { in Sum565() 883 const uint16x8_t sum = (offset == 0) ? Sum3WLo16(ma5) : Sum3WHi16(ma5); in Sum565() 887 (offset == 0) ? vget_low_u8(ma5[1]) : vget_high_u8(ma5[1])); in Sum565() 1357 uint16x8_t b3[2][3], uint8x16_t* const ma5, uint16x8_t* const b5) { in BoxFilterPreProcessLo() argument 1397 CalculateIntermediate5<0>(s5, sq5, scales[0], ma5, b5); in BoxFilterPreProcessLo() 1405 uint16x8_t b3[2][3], uint8x16_t ma5[2], uint16x8_t b5[2]) { in BoxFilterPreProcess() 1446 CalculateIntermediate5<8>(s5[0], sq5, scales[0], &ma5[0], &b5[0]); in BoxFilterPreProcess() 1482 CalculateIntermediate5<0>(s5[1], sq5, scales[0], &ma5[1], &b5[1]); in BoxFilterPreProcess() 1490 uint8x16_t* const ma5, uint16x8_t* const b3, uint16x8_t* const b5) { in BoxFilterPreProcessLastRowLo() argument 1508 CalculateIntermediate5<0>(s5, sq5, scales[0], ma5, b5); in BoxFilterPreProcessLastRowLo() [all …]
|