diff options
Diffstat (limited to 'third_party/aom/av1/common/x86/av1_fwd_txfm1d_sse4.c')
-rw-r--r-- | third_party/aom/av1/common/x86/av1_fwd_txfm1d_sse4.c | 28 |
1 files changed, 14 insertions, 14 deletions
diff --git a/third_party/aom/av1/common/x86/av1_fwd_txfm1d_sse4.c b/third_party/aom/av1/common/x86/av1_fwd_txfm1d_sse4.c index d04b667f1..97d2e74b1 100644 --- a/third_party/aom/av1/common/x86/av1_fwd_txfm1d_sse4.c +++ b/third_party/aom/av1/common/x86/av1_fwd_txfm1d_sse4.c @@ -57,7 +57,7 @@ void av1_fdct32_new_sse4_1(const __m128i *input, __m128i *output, // stage 2 stage_idx++; bit = cos_bit[stage_idx]; - cospi = cospi_arr[bit - cos_bit_min]; + cospi = cospi_arr(bit); buf0[0] = _mm_add_epi32(buf1[0], buf1[15]); buf0[15] = _mm_sub_epi32(buf1[0], buf1[15]); buf0[1] = _mm_add_epi32(buf1[1], buf1[14]); @@ -94,7 +94,7 @@ void av1_fdct32_new_sse4_1(const __m128i *input, __m128i *output, // stage 3 stage_idx++; bit = cos_bit[stage_idx]; - cospi = cospi_arr[bit - cos_bit_min]; + cospi = cospi_arr(bit); buf1[0] = _mm_add_epi32(buf0[0], buf0[7]); buf1[7] = _mm_sub_epi32(buf0[0], buf0[7]); buf1[1] = _mm_add_epi32(buf0[1], buf0[6]); @@ -131,7 +131,7 @@ void av1_fdct32_new_sse4_1(const __m128i *input, __m128i *output, // stage 4 stage_idx++; bit = cos_bit[stage_idx]; - cospi = cospi_arr[bit - cos_bit_min]; + cospi = cospi_arr(bit); buf0[0] = _mm_add_epi32(buf1[0], buf1[3]); buf0[3] = _mm_sub_epi32(buf1[0], buf1[3]); buf0[1] = _mm_add_epi32(buf1[1], buf1[2]); @@ -168,7 +168,7 @@ void av1_fdct32_new_sse4_1(const __m128i *input, __m128i *output, // stage 5 stage_idx++; bit = cos_bit[stage_idx]; - cospi = cospi_arr[bit - cos_bit_min]; + cospi = cospi_arr(bit); btf_32_sse4_1_type0(cospi[32], cospi[32], buf0[0], buf0[1], buf1[0], buf1[1], bit); btf_32_sse4_1_type1(cospi[48], cospi[16], buf0[2], buf0[3], buf1[2], @@ -205,7 +205,7 @@ void av1_fdct32_new_sse4_1(const __m128i *input, __m128i *output, // stage 6 stage_idx++; bit = cos_bit[stage_idx]; - cospi = cospi_arr[bit - cos_bit_min]; + cospi = cospi_arr(bit); buf0[0] = buf1[0]; buf0[1] = buf1[1]; buf0[2] = buf1[2]; @@ -242,7 +242,7 @@ void av1_fdct32_new_sse4_1(const __m128i *input, __m128i *output, // stage 7 stage_idx++; bit = cos_bit[stage_idx]; - cospi = cospi_arr[bit - cos_bit_min]; + cospi = cospi_arr(bit); buf1[0] = buf0[0]; buf1[1] = buf0[1]; buf1[2] = buf0[2]; @@ -279,7 +279,7 @@ void av1_fdct32_new_sse4_1(const __m128i *input, __m128i *output, // stage 8 stage_idx++; bit = cos_bit[stage_idx]; - cospi = cospi_arr[bit - cos_bit_min]; + cospi = cospi_arr(bit); buf0[0] = buf1[0]; buf0[1] = buf1[1]; buf0[2] = buf1[2]; @@ -383,7 +383,7 @@ void av1_fadst4_new_sse4_1(const __m128i *input, __m128i *output, // stage 2 stage_idx++; bit = cos_bit[stage_idx]; - cospi = cospi_arr[bit - cos_bit_min]; + cospi = cospi_arr(bit); btf_32_sse4_1_type0(cospi[8], cospi[56], buf1[0], buf1[1], buf0[0], buf0[1], bit); btf_32_sse4_1_type0(cospi[40], cospi[24], buf1[2], buf1[3], buf0[2], @@ -399,7 +399,7 @@ void av1_fadst4_new_sse4_1(const __m128i *input, __m128i *output, // stage 4 stage_idx++; bit = cos_bit[stage_idx]; - cospi = cospi_arr[bit - cos_bit_min]; + cospi = cospi_arr(bit); buf0[0] = buf1[0]; buf0[1] = buf1[1]; btf_32_sse4_1_type0(cospi[32], cospi[32], buf1[2], buf1[3], buf0[2], @@ -475,7 +475,7 @@ void av1_fadst32_new_sse4_1(const __m128i *input, __m128i *output, // stage 2 stage_idx++; bit = cos_bit[stage_idx]; - cospi = cospi_arr[bit - cos_bit_min]; + cospi = cospi_arr(bit); btf_32_sse4_1_type0(cospi[1], cospi[63], buf1[0], buf1[1], buf0[0], buf0[1], bit); btf_32_sse4_1_type0(cospi[5], cospi[59], buf1[2], buf1[3], buf0[2], buf0[3], @@ -547,7 +547,7 @@ void av1_fadst32_new_sse4_1(const __m128i *input, __m128i *output, // stage 4 stage_idx++; bit = cos_bit[stage_idx]; - cospi = cospi_arr[bit - cos_bit_min]; + cospi = cospi_arr(bit); buf0[0] = buf1[0]; buf0[1] = buf1[1]; buf0[2] = buf1[2]; @@ -619,7 +619,7 @@ void av1_fadst32_new_sse4_1(const __m128i *input, __m128i *output, // stage 6 stage_idx++; bit = cos_bit[stage_idx]; - cospi = cospi_arr[bit - cos_bit_min]; + cospi = cospi_arr(bit); buf0[0] = buf1[0]; buf0[1] = buf1[1]; buf0[2] = buf1[2]; @@ -691,7 +691,7 @@ void av1_fadst32_new_sse4_1(const __m128i *input, __m128i *output, // stage 8 stage_idx++; bit = cos_bit[stage_idx]; - cospi = cospi_arr[bit - cos_bit_min]; + cospi = cospi_arr(bit); buf0[0] = buf1[0]; buf0[1] = buf1[1]; buf0[2] = buf1[2]; @@ -763,7 +763,7 @@ void av1_fadst32_new_sse4_1(const __m128i *input, __m128i *output, // stage 10 stage_idx++; bit = cos_bit[stage_idx]; - cospi = cospi_arr[bit - cos_bit_min]; + cospi = cospi_arr(bit); buf0[0] = buf1[0]; buf0[1] = buf1[1]; btf_32_sse4_1_type0(cospi[32], cospi[32], buf1[2], buf1[3], buf0[2], |