diff --git a/vpx_dsp/x86/variance_avx2.c b/vpx_dsp/x86/variance_avx2.c index 7f6e1343a..087dfab96 100644 --- a/vpx_dsp/x86/variance_avx2.c +++ b/vpx_dsp/x86/variance_avx2.c @@ -35,115 +35,76 @@ DECLARE_ALIGNED(32, static const uint8_t, bilinear_filters_avx2[512]) = { void vpx_get16x16var_avx2(const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int recon_stride, - unsigned int *SSE, int *Sum) { - __m256i src, src_expand_low, src_expand_high, ref, ref_expand_low; - __m256i ref_expand_high, madd_low, madd_high; + unsigned int *sse, int *sum) { unsigned int i, src_2strides, ref_2strides; - __m256i zero_reg = _mm256_setzero_si256(); __m256i sum_ref_src = _mm256_setzero_si256(); __m256i madd_ref_src = _mm256_setzero_si256(); - - // processing two strides in a 256 bit register reducing the number - // of loop stride by half (comparing to the sse2 code) + // process two 16 byte locations in a 256 bit register src_2strides = source_stride << 1; ref_2strides = recon_stride << 1; - for (i = 0; i < 8; i++) { - src = _mm256_castsi128_si256(_mm_loadu_si128((__m128i const *)(src_ptr))); - src = _mm256_inserti128_si256( - src, _mm_loadu_si128((__m128i const *)(src_ptr + source_stride)), 1); - - ref = _mm256_castsi128_si256(_mm_loadu_si128((__m128i const *)(ref_ptr))); - ref = _mm256_inserti128_si256( - ref, _mm_loadu_si128((__m128i const *)(ref_ptr + recon_stride)), 1); - - // expanding to 16 bit each lane - src_expand_low = _mm256_unpacklo_epi8(src, zero_reg); - src_expand_high = _mm256_unpackhi_epi8(src, zero_reg); - - ref_expand_low = _mm256_unpacklo_epi8(ref, zero_reg); - ref_expand_high = _mm256_unpackhi_epi8(ref, zero_reg); - - // src-ref - src_expand_low = _mm256_sub_epi16(src_expand_low, ref_expand_low); - src_expand_high = _mm256_sub_epi16(src_expand_high, ref_expand_high); - - // madd low (src - ref) - madd_low = _mm256_madd_epi16(src_expand_low, src_expand_low); - - // add high to low - src_expand_low = _mm256_add_epi16(src_expand_low, src_expand_high); - - // madd high (src - ref) - madd_high = _mm256_madd_epi16(src_expand_high, src_expand_high); - - sum_ref_src = _mm256_add_epi16(sum_ref_src, src_expand_low); + for (i = 0; i < 8; ++i) { + const __m256i zero_reg = _mm256_setzero_si256(); + const __m256i src0 = + _mm256_castsi128_si256(_mm_loadu_si128((__m128i const *)(src_ptr))); + const __m256i src = _mm256_inserti128_si256( + src0, _mm_loadu_si128((__m128i const *)(src_ptr + source_stride)), 1); + const __m256i ref0 = + _mm256_castsi128_si256(_mm_loadu_si128((__m128i const *)(ref_ptr))); + const __m256i ref = _mm256_inserti128_si256( + ref0, _mm_loadu_si128((__m128i const *)(ref_ptr + recon_stride)), 1); + const __m256i src_lo = _mm256_unpacklo_epi8(src, zero_reg); + const __m256i src_hi = _mm256_unpackhi_epi8(src, zero_reg); + const __m256i ref_lo = _mm256_unpacklo_epi8(ref, zero_reg); + const __m256i ref_hi = _mm256_unpackhi_epi8(ref, zero_reg); + const __m256i diff_lo = _mm256_sub_epi16(src_lo, ref_lo); + const __m256i diff_hi = _mm256_sub_epi16(src_hi, ref_hi); + const __m256i madd_lo = _mm256_madd_epi16(diff_lo, diff_lo); + const __m256i madd_hi = _mm256_madd_epi16(diff_hi, diff_hi); + const __m256i src_ref_diff_sum = _mm256_add_epi16(diff_lo, diff_hi); + sum_ref_src = _mm256_add_epi16(sum_ref_src, src_ref_diff_sum); // add high to low madd_ref_src = - _mm256_add_epi32(madd_ref_src, _mm256_add_epi32(madd_low, madd_high)); + _mm256_add_epi32(madd_ref_src, _mm256_add_epi32(madd_lo, madd_hi)); src_ptr += src_2strides; ref_ptr += ref_2strides; } { - __m128i sum_res, madd_res; - __m128i expand_sum_low, expand_sum_high, expand_sum; - __m128i expand_madd_low, expand_madd_high, expand_madd; - __m128i ex_expand_sum_low, ex_expand_sum_high, ex_expand_sum; - + const __m128i zero_reg = _mm_setzero_si128(); // extract the low lane and add it to the high lane - sum_res = _mm_add_epi16(_mm256_castsi256_si128(sum_ref_src), - _mm256_extractf128_si256(sum_ref_src, 1)); - - madd_res = _mm_add_epi32(_mm256_castsi256_si128(madd_ref_src), - _mm256_extractf128_si256(madd_ref_src, 1)); - - // padding each 2 bytes with another 2 zeroed bytes - expand_sum_low = - _mm_unpacklo_epi16(_mm256_castsi256_si128(zero_reg), sum_res); - expand_sum_high = - _mm_unpackhi_epi16(_mm256_castsi256_si128(zero_reg), sum_res); - - // shifting the sign 16 bits right - expand_sum_low = _mm_srai_epi32(expand_sum_low, 16); - expand_sum_high = _mm_srai_epi32(expand_sum_high, 16); - - expand_sum = _mm_add_epi32(expand_sum_low, expand_sum_high); - - // expand each 32 bits of the madd result to 64 bits - expand_madd_low = - _mm_unpacklo_epi32(madd_res, _mm256_castsi256_si128(zero_reg)); - expand_madd_high = - _mm_unpackhi_epi32(madd_res, _mm256_castsi256_si128(zero_reg)); - - expand_madd = _mm_add_epi32(expand_madd_low, expand_madd_high); - - ex_expand_sum_low = - _mm_unpacklo_epi32(expand_sum, _mm256_castsi256_si128(zero_reg)); - ex_expand_sum_high = - _mm_unpackhi_epi32(expand_sum, _mm256_castsi256_si128(zero_reg)); - - ex_expand_sum = _mm_add_epi32(ex_expand_sum_low, ex_expand_sum_high); - - // shift 8 bytes eight - madd_res = _mm_srli_si128(expand_madd, 8); - sum_res = _mm_srli_si128(ex_expand_sum, 8); - - madd_res = _mm_add_epi32(madd_res, expand_madd); - sum_res = _mm_add_epi32(sum_res, ex_expand_sum); - - *((int *)SSE) = _mm_cvtsi128_si32(madd_res); - - *((int *)Sum) = _mm_cvtsi128_si32(sum_res); + const __m128i sum_ref_src_128 = + _mm_add_epi16(_mm256_castsi256_si128(sum_ref_src), + _mm256_extractf128_si256(sum_ref_src, 1)); + const __m128i madd_ref_src_128 = + _mm_add_epi32(_mm256_castsi256_si128(madd_ref_src), + _mm256_extractf128_si256(madd_ref_src, 1)); + // 16 -> 32 sign extended + const __m128i sum_lo = + _mm_srai_epi32(_mm_unpacklo_epi16(zero_reg, sum_ref_src_128), 16); + // 16 -> 32 sign extended + const __m128i sum_hi = + _mm_srai_epi32(_mm_unpackhi_epi16(zero_reg, sum_ref_src_128), 16); + const __m128i sum_hl = _mm_add_epi32(sum_lo, sum_hi); + const __m128i madd_lo = _mm_unpacklo_epi32(madd_ref_src_128, zero_reg); + const __m128i madd_hi = _mm_unpackhi_epi32(madd_ref_src_128, zero_reg); + const __m128i madd = _mm_add_epi32(madd_lo, madd_hi); + const __m128i ex_sum_lo = _mm_unpacklo_epi32(sum_hl, zero_reg); + const __m128i ex_sum_hi = _mm_unpackhi_epi32(sum_hl, zero_reg); + const __m128i ex_sum = _mm_add_epi32(ex_sum_lo, ex_sum_hi); + *((int *)sse) = + _mm_cvtsi128_si32(_mm_add_epi32(madd, _mm_srli_si128(madd, 8))); + *((int *)sum) = + _mm_cvtsi128_si32(_mm_add_epi32(ex_sum, _mm_srli_si128(ex_sum, 8))); } } static void get32x32var_avx2(const unsigned char *src_ptr, int source_stride, const unsigned char *ref_ptr, int recon_stride, - unsigned int *SSE, int *Sum) { - __m256i src, src_expand_low, src_expand_high, ref, ref_expand_low; - __m256i ref_expand_high, madd_low, madd_high; + unsigned int *sse, int *sum) { + __m256i src, src_expand_lo, src_expand_hi, ref, ref_expand_lo; + __m256i ref_expand_hi, madd_lo, madd_hi; unsigned int i; __m256i zero_reg = _mm256_setzero_si256(); __m256i sum_ref_src = _mm256_setzero_si256(); @@ -156,30 +117,30 @@ static void get32x32var_avx2(const unsigned char *src_ptr, int source_stride, ref = _mm256_loadu_si256((__m256i const *)(ref_ptr)); // expanding to 16 bit each lane - src_expand_low = _mm256_unpacklo_epi8(src, zero_reg); - src_expand_high = _mm256_unpackhi_epi8(src, zero_reg); + src_expand_lo = _mm256_unpacklo_epi8(src, zero_reg); + src_expand_hi = _mm256_unpackhi_epi8(src, zero_reg); - ref_expand_low = _mm256_unpacklo_epi8(ref, zero_reg); - ref_expand_high = _mm256_unpackhi_epi8(ref, zero_reg); + ref_expand_lo = _mm256_unpacklo_epi8(ref, zero_reg); + ref_expand_hi = _mm256_unpackhi_epi8(ref, zero_reg); // src-ref - src_expand_low = _mm256_sub_epi16(src_expand_low, ref_expand_low); - src_expand_high = _mm256_sub_epi16(src_expand_high, ref_expand_high); + src_expand_lo = _mm256_sub_epi16(src_expand_lo, ref_expand_lo); + src_expand_hi = _mm256_sub_epi16(src_expand_hi, ref_expand_hi); // madd low (src - ref) - madd_low = _mm256_madd_epi16(src_expand_low, src_expand_low); + madd_lo = _mm256_madd_epi16(src_expand_lo, src_expand_lo); // add high to low - src_expand_low = _mm256_add_epi16(src_expand_low, src_expand_high); + src_expand_lo = _mm256_add_epi16(src_expand_lo, src_expand_hi); // madd high (src - ref) - madd_high = _mm256_madd_epi16(src_expand_high, src_expand_high); + madd_hi = _mm256_madd_epi16(src_expand_hi, src_expand_hi); - sum_ref_src = _mm256_add_epi16(sum_ref_src, src_expand_low); + sum_ref_src = _mm256_add_epi16(sum_ref_src, src_expand_lo); // add high to low madd_ref_src = - _mm256_add_epi32(madd_ref_src, _mm256_add_epi32(madd_low, madd_high)); + _mm256_add_epi32(madd_ref_src, _mm256_add_epi32(madd_lo, madd_hi)); src_ptr += source_stride; ref_ptr += recon_stride; @@ -187,7 +148,7 @@ static void get32x32var_avx2(const unsigned char *src_ptr, int source_stride, { __m256i expand_sum_low, expand_sum_high, expand_sum; - __m256i expand_madd_low, expand_madd_high, expand_madd; + __m256i expand_madd_lo, expand_madd_hi, expand_madd; __m256i ex_expand_sum_low, ex_expand_sum_high, ex_expand_sum; // padding each 2 bytes with another 2 zeroed bytes @@ -201,10 +162,10 @@ static void get32x32var_avx2(const unsigned char *src_ptr, int source_stride, expand_sum = _mm256_add_epi32(expand_sum_low, expand_sum_high); // expand each 32 bits of the madd result to 64 bits - expand_madd_low = _mm256_unpacklo_epi32(madd_ref_src, zero_reg); - expand_madd_high = _mm256_unpackhi_epi32(madd_ref_src, zero_reg); + expand_madd_lo = _mm256_unpacklo_epi32(madd_ref_src, zero_reg); + expand_madd_hi = _mm256_unpackhi_epi32(madd_ref_src, zero_reg); - expand_madd = _mm256_add_epi32(expand_madd_low, expand_madd_high); + expand_madd = _mm256_add_epi32(expand_madd_lo, expand_madd_hi); ex_expand_sum_low = _mm256_unpacklo_epi32(expand_sum, zero_reg); ex_expand_sum_high = _mm256_unpackhi_epi32(expand_sum, zero_reg); @@ -219,11 +180,11 @@ static void get32x32var_avx2(const unsigned char *src_ptr, int source_stride, sum_ref_src = _mm256_add_epi32(sum_ref_src, ex_expand_sum); // extract the low lane and the high lane and add the results - *((int *)SSE) = + *((int *)sse) = _mm_cvtsi128_si32(_mm256_castsi256_si128(madd_ref_src)) + _mm_cvtsi128_si32(_mm256_extractf128_si256(madd_ref_src, 1)); - *((int *)Sum) = _mm_cvtsi128_si32(_mm256_castsi256_si128(sum_ref_src)) + + *((int *)sum) = _mm_cvtsi128_si32(_mm256_castsi256_si128(sum_ref_src)) + _mm_cvtsi128_si32(_mm256_extractf128_si256(sum_ref_src, 1)); } }