/external/libgav1/libgav1/src/dsp/x86/ |
D | loop_restoration_10bit_sse4.cc | 1214 __m128i sum_b343[2], __m128i sum_b444[2], in Store343_444() 1220 sum_b343[0] = _mm_sub_epi32(sum_b444[0], sum_b111[0]); in Store343_444() 1221 sum_b343[0] = _mm_add_epi32(sum_b343[0], b[1]); in Store343_444() 1225 sum_b343[1] = _mm_sub_epi32(sum_b444[1], sum_b111[1]); in Store343_444() 1226 sum_b343[1] = _mm_add_epi32(sum_b343[1], b[1]); in Store343_444() 1228 StoreAligned32U32(b343 + x, sum_b343); in Store343_444() 1233 __m128i* const sum_ma444, __m128i sum_b343[2], in Store343_444Lo() 1243 Store343_444(b3, x, sum_b343, sum_b444, b343, b444); in Store343_444Lo() 1248 __m128i* const sum_ma444, __m128i sum_b343[2], in Store343_444Hi() 1258 Store343_444(b3, x, sum_b343, sum_b444, b343, b444); in Store343_444Hi() [all …]
|
D | loop_restoration_sse4.cc | 1353 __m128i sum_b343[2], __m128i sum_b444[2], in Store343_444() 1362 sum_b343[0] = _mm_sub_epi32(sum_b444[0], sum_b111[0]); in Store343_444() 1363 sum_b343[1] = _mm_sub_epi32(sum_b444[1], sum_b111[1]); in Store343_444() 1364 sum_b343[0] = VaddwLo16(sum_b343[0], b[1]); in Store343_444() 1365 sum_b343[1] = VaddwHi16(sum_b343[1], b[1]); in Store343_444() 1366 StoreAligned32U32(b343 + x, sum_b343); in Store343_444() 1371 __m128i* const sum_ma444, __m128i sum_b343[2], in Store343_444Lo() 1381 Store343_444(b3, x, sum_b343, sum_b444, b343, b444); in Store343_444Lo() 1386 __m128i* const sum_ma444, __m128i sum_b343[2], in Store343_444Hi() 1396 Store343_444(b3, x, sum_b343, sum_b444, b343, b444); in Store343_444Hi() [all …]
|
D | loop_restoration_10bit_avx2.cc | 1663 __m256i sum_b343[2], __m256i sum_b444[2], in Store343_444() 1669 sum_b343[0] = _mm256_sub_epi32(sum_b444[0], sum_b111[0]); in Store343_444() 1670 sum_b343[0] = _mm256_add_epi32(sum_b343[0], b[1]); in Store343_444() 1674 sum_b343[1] = _mm256_sub_epi32(sum_b444[1], sum_b111[1]); in Store343_444() 1675 sum_b343[1] = _mm256_add_epi32(sum_b343[1], b[1]); in Store343_444() 1677 StoreAligned64(b343 + x, sum_b343); in Store343_444() 1682 __m256i* const sum_ma444, __m256i sum_b343[2], in Store343_444Lo() 1692 Store343_444(b3, x, sum_b343, sum_b444, b343, b444); in Store343_444Lo() 1697 __m256i* const sum_ma444, __m256i sum_b343[2], in Store343_444Hi() 1707 Store343_444(b3, x + kMaStoreOffset, sum_b343, sum_b444, b343, b444); in Store343_444Hi() [all …]
|
D | loop_restoration_avx2.cc | 1664 __m256i sum_b343[2], __m256i sum_b444[2], in Store343_444() 1673 sum_b343[0] = _mm256_sub_epi32(sum_b444[0], sum_b111[0]); in Store343_444() 1674 sum_b343[1] = _mm256_sub_epi32(sum_b444[1], sum_b111[1]); in Store343_444() 1675 sum_b343[0] = VaddwLo16(sum_b343[0], b[1]); in Store343_444() 1676 sum_b343[1] = VaddwHi16(sum_b343[1], b[1]); in Store343_444() 1677 StoreAligned64(b343 + x, sum_b343); in Store343_444() 1682 __m256i* const sum_ma444, __m256i sum_b343[2], in Store343_444Lo() 1692 Store343_444(b3, x, sum_b343, sum_b444, b343, b444); in Store343_444Lo() 1697 __m256i* const sum_ma444, __m256i sum_b343[2], in Store343_444Hi() 1707 Store343_444(b3, x, sum_b343, sum_b444, b343, b444); in Store343_444Hi() [all …]
|
/external/libgav1/libgav1/src/dsp/arm/ |
D | loop_restoration_neon.cc | 1094 uint32x4x2_t* const sum_b343, in Store343_444() argument 1110 sum_b343->val[0] = vsubq_u32(sum_b444->val[0], sum_b111.val[0]); in Store343_444() 1111 sum_b343->val[1] = vsubq_u32(sum_b444->val[1], sum_b111.val[1]); in Store343_444() 1112 sum_b343->val[0] = vaddw_u16(sum_b343->val[0], low[1]); in Store343_444() 1113 sum_b343->val[1] = vaddw_u16(sum_b343->val[1], high[1]); in Store343_444() 1116 vst1q_u32(b343 + x + 0, sum_b343->val[0]); in Store343_444() 1117 vst1q_u32(b343 + x + 4, sum_b343->val[1]); in Store343_444() 1125 uint32x4x2_t* const sum_b343, uint16_t* const ma343, in Store343_444() argument 1130 Store343_444<offset>(ma3, b3, x, sum_ma343, &sum_ma444, sum_b343, &sum_b444, in Store343_444() 1140 uint32x4x2_t sum_b343; in Store343_444() local [all …]
|