vpx_subpixel_8t_intrin_ssse3.c | Explore in Territory

/*
 *  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
 *
 *  Use of this source code is governed by a BSD-style license
 *  that can be found in the LICENSE file in the root of the source
 *  tree. An additional intellectual property rights grant can be found
 *  in the file PATENTS.  All contributing project authors may
 *  be found in the AUTHORS file in the root of the source tree.
 */

#include <tmmintrin.h>  // SSSE3

#include <string.h>

#include "./vpx_config.h"
#include "./vpx_dsp_rtcd.h"
#include "vpx_dsp/vpx_filter.h"
#include "vpx_dsp/x86/convolve.h"
#include "vpx_dsp/x86/convolve_sse2.h"
#include "vpx_dsp/x86/convolve_ssse3.h"
#include "vpx_dsp/x86/mem_sse2.h"
#include "vpx_dsp/x86/transpose_sse2.h"
#include "vpx_mem/vpx_mem.h"
#include "vpx_ports/mem.h"

static INLINE __m128i shuffle_filter_convolve8_8_ssse3(
    const __m128i *const s, const int16_t *const filter) { … }

// Used by the avx2 implementation.
#if VPX_ARCH_X86_64
// Use the intrinsics below
filter8_1dfunction vpx_filter_block1d4_h8_intrin_ssse3;
filter8_1dfunction vpx_filter_block1d8_h8_intrin_ssse3;
filter8_1dfunction vpx_filter_block1d8_v8_intrin_ssse3;
#define vpx_filter_block1d4_h8_ssse3 …
#define vpx_filter_block1d8_h8_ssse3 …
#define vpx_filter_block1d8_v8_ssse3 …
#else  // VPX_ARCH_X86
// Use the assembly in vpx_dsp/x86/vpx_subpixel_8t_ssse3.asm.
filter8_1dfunction vpx_filter_block1d4_h8_ssse3;
filter8_1dfunction vpx_filter_block1d8_h8_ssse3;
filter8_1dfunction vpx_filter_block1d8_v8_ssse3;
#endif

#if VPX_ARCH_X86_64
void vpx_filter_block1d4_h8_intrin_ssse3(
    const uint8_t *src_ptr, ptrdiff_t src_pitch, uint8_t *output_ptr,
    ptrdiff_t output_pitch, uint32_t output_height, const int16_t *filter) { … }

void vpx_filter_block1d8_h8_intrin_ssse3(
    const uint8_t *src_ptr, ptrdiff_t src_pitch, uint8_t *output_ptr,
    ptrdiff_t output_pitch, uint32_t output_height, const int16_t *filter) { … }

void vpx_filter_block1d8_v8_intrin_ssse3(
    const uint8_t *src_ptr, ptrdiff_t src_pitch, uint8_t *output_ptr,
    ptrdiff_t out_pitch, uint32_t output_height, const int16_t *filter) { … }
#endif  // VPX_ARCH_X86_64

static void vpx_filter_block1d16_h4_ssse3(const uint8_t *src_ptr,
                                          ptrdiff_t src_stride,
                                          uint8_t *dst_ptr,
                                          ptrdiff_t dst_stride, uint32_t height,
                                          const int16_t *kernel) { … }

static void vpx_filter_block1d16_v4_ssse3(const uint8_t *src_ptr,
                                          ptrdiff_t src_stride,
                                          uint8_t *dst_ptr,
                                          ptrdiff_t dst_stride, uint32_t height,
                                          const int16_t *kernel) { … }

static void vpx_filter_block1d8_h4_ssse3(const uint8_t *src_ptr,
                                         ptrdiff_t src_stride, uint8_t *dst_ptr,
                                         ptrdiff_t dst_stride, uint32_t height,
                                         const int16_t *kernel) { … }

static void vpx_filter_block1d8_v4_ssse3(const uint8_t *src_ptr,
                                         ptrdiff_t src_stride, uint8_t *dst_ptr,
                                         ptrdiff_t dst_stride, uint32_t height,
                                         const int16_t *kernel) { … }

static void vpx_filter_block1d4_h4_ssse3(const uint8_t *src_ptr,
                                         ptrdiff_t src_stride, uint8_t *dst_ptr,
                                         ptrdiff_t dst_stride, uint32_t height,
                                         const int16_t *kernel) { … }

static void vpx_filter_block1d4_v4_ssse3(const uint8_t *src_ptr,
                                         ptrdiff_t src_stride, uint8_t *dst_ptr,
                                         ptrdiff_t dst_stride, uint32_t height,
                                         const int16_t *kernel) { … }

// From vpx_dsp/x86/vpx_subpixel_8t_ssse3.asm
filter8_1dfunction vpx_filter_block1d16_v8_ssse3;
filter8_1dfunction vpx_filter_block1d16_h8_ssse3;
filter8_1dfunction vpx_filter_block1d4_v8_ssse3;
filter8_1dfunction vpx_filter_block1d16_v8_avg_ssse3;
filter8_1dfunction vpx_filter_block1d16_h8_avg_ssse3;
filter8_1dfunction vpx_filter_block1d8_v8_avg_ssse3;
filter8_1dfunction vpx_filter_block1d8_h8_avg_ssse3;
filter8_1dfunction vpx_filter_block1d4_v8_avg_ssse3;
filter8_1dfunction vpx_filter_block1d4_h8_avg_ssse3;

// Use the [vh]8 version because there is no [vh]4 implementation.
#define vpx_filter_block1d16_v4_avg_ssse3 …
#define vpx_filter_block1d16_h4_avg_ssse3 …
#define vpx_filter_block1d8_v4_avg_ssse3 …
#define vpx_filter_block1d8_h4_avg_ssse3 …
#define vpx_filter_block1d4_v4_avg_ssse3 …
#define vpx_filter_block1d4_h4_avg_ssse3 …

// From vpx_dsp/x86/vpx_subpixel_bilinear_ssse3.asm
filter8_1dfunction vpx_filter_block1d16_v2_ssse3;
filter8_1dfunction vpx_filter_block1d16_h2_ssse3;
filter8_1dfunction vpx_filter_block1d8_v2_ssse3;
filter8_1dfunction vpx_filter_block1d8_h2_ssse3;
filter8_1dfunction vpx_filter_block1d4_v2_ssse3;
filter8_1dfunction vpx_filter_block1d4_h2_ssse3;
filter8_1dfunction vpx_filter_block1d16_v2_avg_ssse3;
filter8_1dfunction vpx_filter_block1d16_h2_avg_ssse3;
filter8_1dfunction vpx_filter_block1d8_v2_avg_ssse3;
filter8_1dfunction vpx_filter_block1d8_h2_avg_ssse3;
filter8_1dfunction vpx_filter_block1d4_v2_avg_ssse3;
filter8_1dfunction vpx_filter_block1d4_h2_avg_ssse3;

// void vpx_convolve8_horiz_ssse3(const uint8_t *src, ptrdiff_t src_stride,
//                                uint8_t *dst, ptrdiff_t dst_stride,
//                                const InterpKernel *filter, int x0_q4,
//                                int32_t x_step_q4, int y0_q4, int y_step_q4,
//                                int w, int h);
// void vpx_convolve8_vert_ssse3(const uint8_t *src, ptrdiff_t src_stride,
//                               uint8_t *dst, ptrdiff_t dst_stride,
//                               const InterpKernel *filter, int x0_q4,
//                               int32_t x_step_q4, int y0_q4, int y_step_q4,
//                               int w, int h);
// void vpx_convolve8_avg_horiz_ssse3(const uint8_t *src, ptrdiff_t src_stride,
//                                    uint8_t *dst, ptrdiff_t dst_stride,
//                                    const InterpKernel *filter, int x0_q4,
//                                    int32_t x_step_q4, int y0_q4,
//                                    int y_step_q4, int w, int h);
// void vpx_convolve8_avg_vert_ssse3(const uint8_t *src, ptrdiff_t src_stride,
//                                   uint8_t *dst, ptrdiff_t dst_stride,
//                                   const InterpKernel *filter, int x0_q4,
//                                   int32_t x_step_q4, int y0_q4,
//                                   int y_step_q4, int w, int h);
FUN_CONV_1D(…)
FUN_CONV_1D(…)
FUN_CONV_1D(…)
FUN_CONV_1D(…)

static void filter_horiz_w8_ssse3(const uint8_t *const src,
                                  const ptrdiff_t src_stride,
                                  uint8_t *const dst,
                                  const int16_t *const x_filter) { … }

static void transpose8x8_to_dst(const uint8_t *const src,
                                const ptrdiff_t src_stride, uint8_t *const dst,
                                const ptrdiff_t dst_stride) { … }

static void scaledconvolve_horiz_w8(const uint8_t *src,
                                    const ptrdiff_t src_stride, uint8_t *dst,
                                    const ptrdiff_t dst_stride,
                                    const InterpKernel *const x_filters,
                                    const int x0_q4, const int x_step_q4,
                                    const int w, const int h) { … }

static void filter_horiz_w4_ssse3(const uint8_t *const src,
                                  const ptrdiff_t src_stride,
                                  uint8_t *const dst,
                                  const int16_t *const filter) { … }

static void transpose4x4_to_dst(const uint8_t *const src,
                                const ptrdiff_t src_stride, uint8_t *const dst,
                                const ptrdiff_t dst_stride) { … }

static void scaledconvolve_horiz_w4(const uint8_t *src,
                                    const ptrdiff_t src_stride, uint8_t *dst,
                                    const ptrdiff_t dst_stride,
                                    const InterpKernel *const x_filters,
                                    const int x0_q4, const int x_step_q4,
                                    const int w, const int h) { … }

static __m128i filter_vert_kernel(const __m128i *const s,
                                  const int16_t *const filter) { … }

static void filter_vert_w4_ssse3(const uint8_t *const src,
                                 const ptrdiff_t src_stride, uint8_t *const dst,
                                 const int16_t *const filter) { … }

static void scaledconvolve_vert_w4(
    const uint8_t *src, const ptrdiff_t src_stride, uint8_t *const dst,
    const ptrdiff_t dst_stride, const InterpKernel *const y_filters,
    const int y0_q4, const int y_step_q4, const int w, const int h) { … }

static void filter_vert_w8_ssse3(const uint8_t *const src,
                                 const ptrdiff_t src_stride, uint8_t *const dst,
                                 const int16_t *const filter) { … }

static void scaledconvolve_vert_w8(
    const uint8_t *src, const ptrdiff_t src_stride, uint8_t *const dst,
    const ptrdiff_t dst_stride, const InterpKernel *const y_filters,
    const int y0_q4, const int y_step_q4, const int w, const int h) { … }

static void filter_vert_w16_ssse3(const uint8_t *src,
                                  const ptrdiff_t src_stride,
                                  uint8_t *const dst,
                                  const int16_t *const filter, const int w) { … }

static void scaledconvolve_vert_w16(
    const uint8_t *src, const ptrdiff_t src_stride, uint8_t *const dst,
    const ptrdiff_t dst_stride, const InterpKernel *const y_filters,
    const int y0_q4, const int y_step_q4, const int w, const int h) { … }

void vpx_scaled_2d_ssse3(const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst,
                         ptrdiff_t dst_stride, const InterpKernel *filter,
                         int x0_q4, int x_step_q4, int y0_q4, int y_step_q4,
                         int w, int h) { … }

// void vpx_convolve8_ssse3(const uint8_t *src, ptrdiff_t src_stride,
//                          uint8_t *dst, ptrdiff_t dst_stride,
//                          const InterpKernel *filter, int x0_q4,
//                          int32_t x_step_q4, int y0_q4, int y_step_q4,
//                          int w, int h);
// void vpx_convolve8_avg_ssse3(const uint8_t *src, ptrdiff_t src_stride,
//                              uint8_t *dst, ptrdiff_t dst_stride,
//                              const InterpKernel *filter, int x0_q4,
//                              int32_t x_step_q4, int y0_q4, int y_step_q4,
//                              int w, int h);
FUN_CONV_2D(…)
FUN_CONV_2D(…)
chromium/third_party/libvpx/source/libvpx/vpx_dsp/x86/vpx_subpixel_8t_intrin_ssse3.c
