Line data    Source code

       1             : /*
       2             :  * Copyright (c) 2018, Alliance for Open Media. All rights reserved
       3             :  *
       4             :  * This source code is subject to the terms of the BSD 2 Clause License and
       5             :  * the Alliance for Open Media Patent License 1.0. If the BSD 2 Clause License
       6             :  * was not distributed with this source code in the LICENSE file, you can
       7             :  * obtain it at www.aomedia.org/license/software. If the Alliance for Open
       8             :  * Media Patent License 1.0 was not distributed with this source code in the
       9             :  * PATENTS file, you can obtain it at www.aomedia.org/license/patent.
      10             :  */
      11             : 
      12             : #include <immintrin.h>
      13             : 
      14             : #include "EbDefinitions.h"
      15             : #include "fft_common.h"
      16             : 
      17             : extern void eb_aom_transpose_float_sse2(const float *A, float *B, int32_t n);
      18             : extern void eb_aom_fft_unpack_2d_output_sse2(const float *col_fft, float *output,
      19             :     int32_t n);
      20             : 
      21             : // Generate the 1d forward transforms for float using _mm256
      22           0 : GEN_FFT_8(static INLINE void, avx2, float, __m256, _mm256_load_ps,
      23             :     _mm256_store_ps, _mm256_set1_ps, _mm256_add_ps, _mm256_sub_ps,
      24             :     _mm256_mul_ps);
      25           0 : GEN_FFT_16(static INLINE void, avx2, float, __m256, _mm256_load_ps,
      26             :     _mm256_store_ps, _mm256_set1_ps, _mm256_add_ps, _mm256_sub_ps,
      27             :     _mm256_mul_ps);
      28           0 : GEN_FFT_32(static INLINE void, avx2, float, __m256, _mm256_load_ps,
      29             :     _mm256_store_ps, _mm256_set1_ps, _mm256_add_ps, _mm256_sub_ps,
      30             :     _mm256_mul_ps);
      31             : 
      32           0 : void eb_aom_fft8x8_float_avx2(const float *input, float *temp, float *output) {
      33           0 :     eb_aom_fft_2d_gen(input, temp, output, 8, eb_aom_fft1d_8_avx2,
      34             :         eb_aom_transpose_float_sse2, eb_aom_fft_unpack_2d_output_sse2, 8);
      35           0 : }
      36             : 
      37           0 : void eb_aom_fft16x16_float_avx2(const float *input, float *temp, float *output) {
      38           0 :     eb_aom_fft_2d_gen(input, temp, output, 16, eb_aom_fft1d_16_avx2,
      39             :         eb_aom_transpose_float_sse2, eb_aom_fft_unpack_2d_output_sse2, 8);
      40           0 : }
      41             : 
      42           0 : void eb_aom_fft32x32_float_avx2(const float *input, float *temp, float *output) {
      43           0 :     eb_aom_fft_2d_gen(input, temp, output, 32, eb_aom_fft1d_32_avx2,
      44             :         eb_aom_transpose_float_sse2, eb_aom_fft_unpack_2d_output_sse2, 8);
      45           0 : }
      46             : 
      47             : // Generate the 1d inverse transforms for float using _mm256
      48           0 : GEN_IFFT_8(static INLINE void, avx2, float, __m256, _mm256_load_ps,
      49             :     _mm256_store_ps, _mm256_set1_ps, _mm256_add_ps, _mm256_sub_ps,
      50             :     _mm256_mul_ps);
      51           0 : GEN_IFFT_16(static INLINE void, avx2, float, __m256, _mm256_load_ps,
      52             :     _mm256_store_ps, _mm256_set1_ps, _mm256_add_ps, _mm256_sub_ps,
      53             :     _mm256_mul_ps);
      54           0 : GEN_IFFT_32(static INLINE void, avx2, float, __m256, _mm256_load_ps,
      55             :     _mm256_store_ps, _mm256_set1_ps, _mm256_add_ps, _mm256_sub_ps,
      56             :     _mm256_mul_ps);
      57             : 
      58           0 : void eb_aom_ifft8x8_float_avx2(const float *input, float *temp, float *output) {
      59           0 :     eb_aom_ifft_2d_gen(input, temp, output, 8, eb_aom_fft1d_8_float, eb_aom_fft1d_8_avx2,
      60             :         eb_aom_ifft1d_8_avx2, eb_aom_transpose_float_sse2, 8);
      61           0 : }
      62             : 
      63           0 : void eb_aom_ifft16x16_float_avx2(const float *input, float *temp, float *output) {
      64           0 :     eb_aom_ifft_2d_gen(input, temp, output, 16, eb_aom_fft1d_16_float,
      65             :         eb_aom_fft1d_16_avx2, eb_aom_ifft1d_16_avx2,
      66             :         eb_aom_transpose_float_sse2, 8);
      67           0 : }
      68             : 
      69           0 : void eb_aom_ifft32x32_float_avx2(const float *input, float *temp, float *output) {
      70           0 :     eb_aom_ifft_2d_gen(input, temp, output, 32, eb_aom_fft1d_32_float,
      71             :         eb_aom_fft1d_32_avx2, eb_aom_ifft1d_32_avx2,
      72             :         eb_aom_transpose_float_sse2, 8);
      73           0 : }