| /external/libgav1/libgav1/src/dsp/ |
| D | loop_restoration.cc | 367 uint16_t* const ma565, uint32_t* const b565) { in BoxFilterPreProcess5() argument
381 ma565[x] = Sum565(sgr_buffer->ma + x); in BoxFilterPreProcess5()
427 const uint16_t* const ma565[2], in BoxFilterPass1Kernel()
430 p[0] = CalculateFilteredOutput<Pixel>(src0, ma565[0][x] + ma565[1][x], in BoxFilterPass1Kernel()
432 p[1] = CalculateFilteredOutput<Pixel>(src1, ma565[1][x], b565[1][x], 4); in BoxFilterPass1Kernel()
477 uint16_t* const ma565[2], uint32_t* const b565[2], in BoxFilterPass1()
480 ma565[1], b565[1]); in BoxFilterPass1()
484 BoxFilterPass1Kernel<Pixel>(src[x], src[stride + x], ma565, b565, x, p); in BoxFilterPass1()
520 uint16_t* const ma565[2], uint32_t* const b343[4], in BoxFilter()
524 sgr_buffer, ma565[1], b565[1]); in BoxFilter()
[all …]
|
| D | common.h | 56 alignas(kMaxAlignment) uint16_t ma565[2 * kRestorationUnitWidth]; member
|
| /external/libgav1/libgav1/src/dsp/arm/ |
| D | loop_restoration_neon.cc | 1578 uint16_t* ma565, uint32_t* b565) { in BoxSumFilterPreProcess5() argument
1594 vst1q_u16(ma565, ma[0]); in BoxSumFilterPreProcess5()
1600 vst1q_u16(ma565 + 8, ma[1]); in BoxSumFilterPreProcess5()
1609 ma565 += 16; in BoxSumFilterPreProcess5()
1665 uint16_t* const ma343[4], uint16_t* const ma444, uint16_t* ma565, in BoxSumFilterPreProcess() argument
1698 vst1q_u16(ma565, ma[0]); in BoxSumFilterPreProcess()
1699 vst1q_u16(ma565 + 8, ma[1]); in BoxSumFilterPreProcess()
1714 ma565 += 16; in BoxSumFilterPreProcess()
1800 const int16_t w0, uint16_t* const ma565[2], uint32_t* const b565[2], in BoxFilterPass1()
1818 vst1q_u16(ma565[1] + x, ma[1]); in BoxFilterPass1()
[all …]
|
| /external/libgav1/libgav1/src/dsp/x86/ |
| D | loop_restoration_10bit_sse4.cc | 1578 const ptrdiff_t sum_width, uint16_t* ma565, in BoxSumFilterPreProcess5() argument
1607 StoreAligned32U16(ma565, ma); in BoxSumFilterPreProcess5()
1622 ma565 += 16; in BoxSumFilterPreProcess5()
1683 uint16_t* ma565, uint32_t* const b343[4], uint32_t* const b444, in BoxSumFilterPreProcess() argument
1727 StoreAligned32U16(ma565, ma); in BoxSumFilterPreProcess()
1745 ma565 += 16; in BoxSumFilterPreProcess()
1833 const uint32_t scale, const int16_t w0, uint16_t* const ma565[2], in BoxFilterPass1()
1861 StoreAligned16(ma565[1] + x, ma[1]); in BoxFilterPass1()
1866 ma[0] = LoadAligned16(ma565[0] + x); in BoxFilterPass1()
1874 StoreAligned16(ma565[1] + x + 8, ma[1]); in BoxFilterPass1()
[all …]
|
| D | loop_restoration_sse4.cc | 1712 const ptrdiff_t sum_width, uint16_t* ma565, in BoxSumFilterPreProcess5() argument
1733 StoreAligned32U16(ma565, ma); in BoxSumFilterPreProcess5()
1743 ma565 += 16; in BoxSumFilterPreProcess5()
1797 uint16_t* ma565, uint32_t* const b343[4], uint32_t* const b444, in BoxSumFilterPreProcess() argument
1833 StoreAligned32U16(ma565, ma); in BoxSumFilterPreProcess()
1844 ma565 += 16; in BoxSumFilterPreProcess()
1926 const uint32_t scale, const int16_t w0, uint16_t* const ma565[2], in BoxFilterPass1()
1946 StoreAligned16(ma565[1] + x, ma[1]); in BoxFilterPass1()
1953 ma[0] = LoadAligned16(ma565[0] + x); in BoxFilterPass1()
1961 StoreAligned16(ma565[1] + x + 8, ma[1]); in BoxFilterPass1()
[all …]
|
| D | loop_restoration_10bit_avx2.cc | 2170 const ptrdiff_t sum_width, uint16_t* ma565, in BoxSumFilterPreProcess5() argument
2201 StoreAligned64_ma(ma565, ma); in BoxSumFilterPreProcess5()
2213 ma565 += 32; in BoxSumFilterPreProcess5()
2277 uint16_t* ma565, uint32_t* const b343[4], uint32_t* const b444, in BoxSumFilterPreProcess() argument
2329 StoreAligned64_ma(ma565, ma); in BoxSumFilterPreProcess()
2347 ma565 += 32; in BoxSumFilterPreProcess()
2436 const uint32_t scale, const int16_t w0, uint16_t* const ma565[2], in BoxFilterPass1()
2469 StoreAligned32(ma565[1] + x + 0, ma[1]); in BoxFilterPass1()
2470 StoreAligned32(ma565[1] + x + 16, ma[3]); in BoxFilterPass1()
2476 ma[0] = LoadAligned32(ma565[0] + x); in BoxFilterPass1()
[all …]
|
| D | loop_restoration_avx2.cc | 2059 const ptrdiff_t sum_width, uint16_t* ma565, in BoxSumFilterPreProcess5() argument
2082 StoreAligned64(ma565, ma); in BoxSumFilterPreProcess5()
2091 ma565 += 32; in BoxSumFilterPreProcess5()
2148 uint16_t* ma565, uint32_t* const b343[4], uint32_t* const b444, in BoxSumFilterPreProcess() argument
2188 StoreAligned64(ma565, ma); in BoxSumFilterPreProcess()
2201 ma565 += 32; in BoxSumFilterPreProcess()
2284 const uint32_t scale, const int16_t w0, uint16_t* const ma565[2], in BoxFilterPass1()
2307 StoreAligned64(ma565[1] + x, ma + 1); in BoxFilterPass1()
2316 ma[0] = LoadAligned32(ma565[0] + x); in BoxFilterPass1()
2324 ma[1] = LoadAligned32(ma565[0] + x + 16); in BoxFilterPass1()
[all …]
|