/* * Copyright (c) 2018 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 <emmintrin.h> #include "./vpx_dsp_rtcd.h" #include "vpx/vpx_integer.h" #include "vpx_dsp/x86/convolve.h" #include "vpx_dsp/x86/convolve_sse2.h" #include "vpx_ports/mem.h" #define CONV8_ROUNDING_BITS … #define CONV8_ROUNDING_NUM … static void vpx_filter_block1d16_h4_sse2(const uint8_t *src_ptr, ptrdiff_t src_stride, uint8_t *dst_ptr, ptrdiff_t dst_stride, uint32_t height, const int16_t *kernel) { … } /* The macro used to generate functions shifts the src_ptr up by 3 rows already * */ static void vpx_filter_block1d16_v4_sse2(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_sse2(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_sse2(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_sse2(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_sse2(const uint8_t *src_ptr, ptrdiff_t src_stride, uint8_t *dst_ptr, ptrdiff_t dst_stride, uint32_t height, const int16_t *kernel) { … } #if CONFIG_VP9_HIGHBITDEPTH && VPX_ARCH_X86_64 static void vpx_highbd_filter_block1d4_h4_sse2( const uint16_t *src_ptr, ptrdiff_t src_stride, uint16_t *dst_ptr, ptrdiff_t dst_stride, uint32_t height, const int16_t *kernel, int bd) { … } static void vpx_highbd_filter_block1d4_v4_sse2( const uint16_t *src_ptr, ptrdiff_t src_stride, uint16_t *dst_ptr, ptrdiff_t dst_stride, uint32_t height, const int16_t *kernel, int bd) { … } static void vpx_highbd_filter_block1d8_h4_sse2( const uint16_t *src_ptr, ptrdiff_t src_stride, uint16_t *dst_ptr, ptrdiff_t dst_stride, uint32_t height, const int16_t *kernel, int bd) { … } static void vpx_highbd_filter_block1d8_v4_sse2( const uint16_t *src_ptr, ptrdiff_t src_stride, uint16_t *dst_ptr, ptrdiff_t dst_stride, uint32_t height, const int16_t *kernel, int bd) { … } static void vpx_highbd_filter_block1d16_h4_sse2( const uint16_t *src_ptr, ptrdiff_t src_stride, uint16_t *dst_ptr, ptrdiff_t dst_stride, uint32_t height, const int16_t *kernel, int bd) { … } static void vpx_highbd_filter_block1d16_v4_sse2( const uint16_t *src_ptr, ptrdiff_t src_stride, uint16_t *dst_ptr, ptrdiff_t dst_stride, uint32_t height, const int16_t *kernel, int bd) { … } #endif // CONFIG_VP9_HIGHBITDEPTH && VPX_ARCH_X86_64 // From vpx_subpixel_8t_sse2.asm. filter8_1dfunction vpx_filter_block1d16_v8_sse2; filter8_1dfunction vpx_filter_block1d16_h8_sse2; filter8_1dfunction vpx_filter_block1d8_v8_sse2; filter8_1dfunction vpx_filter_block1d8_h8_sse2; filter8_1dfunction vpx_filter_block1d4_v8_sse2; filter8_1dfunction vpx_filter_block1d4_h8_sse2; filter8_1dfunction vpx_filter_block1d16_v8_avg_sse2; filter8_1dfunction vpx_filter_block1d16_h8_avg_sse2; filter8_1dfunction vpx_filter_block1d8_v8_avg_sse2; filter8_1dfunction vpx_filter_block1d8_h8_avg_sse2; filter8_1dfunction vpx_filter_block1d4_v8_avg_sse2; filter8_1dfunction vpx_filter_block1d4_h8_avg_sse2; // Use the [vh]8 version because there is no [vh]4 implementation. #define vpx_filter_block1d16_v4_avg_sse2 … #define vpx_filter_block1d16_h4_avg_sse2 … #define vpx_filter_block1d8_v4_avg_sse2 … #define vpx_filter_block1d8_h4_avg_sse2 … #define vpx_filter_block1d4_v4_avg_sse2 … #define vpx_filter_block1d4_h4_avg_sse2 … // From vpx_dsp/x86/vpx_subpixel_bilinear_sse2.asm. filter8_1dfunction vpx_filter_block1d16_v2_sse2; filter8_1dfunction vpx_filter_block1d16_h2_sse2; filter8_1dfunction vpx_filter_block1d8_v2_sse2; filter8_1dfunction vpx_filter_block1d8_h2_sse2; filter8_1dfunction vpx_filter_block1d4_v2_sse2; filter8_1dfunction vpx_filter_block1d4_h2_sse2; filter8_1dfunction vpx_filter_block1d16_v2_avg_sse2; filter8_1dfunction vpx_filter_block1d16_h2_avg_sse2; filter8_1dfunction vpx_filter_block1d8_v2_avg_sse2; filter8_1dfunction vpx_filter_block1d8_h2_avg_sse2; filter8_1dfunction vpx_filter_block1d4_v2_avg_sse2; filter8_1dfunction vpx_filter_block1d4_h2_avg_sse2; // void vpx_convolve8_horiz_sse2(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_sse2(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_sse2(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_sse2(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(…) // void vpx_convolve8_sse2(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_sse2(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(…) #if CONFIG_VP9_HIGHBITDEPTH && VPX_ARCH_X86_64 // From vpx_dsp/x86/vpx_high_subpixel_8t_sse2.asm. highbd_filter8_1dfunction vpx_highbd_filter_block1d16_v8_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d16_h8_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d8_v8_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d8_h8_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d4_v8_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d4_h8_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d16_v8_avg_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d16_h8_avg_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d8_v8_avg_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d8_h8_avg_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d4_v8_avg_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d4_h8_avg_sse2; // Use the [vh]8 version because there is no [vh]4 implementation. #define vpx_highbd_filter_block1d16_v4_avg_sse2 … #define vpx_highbd_filter_block1d16_h4_avg_sse2 … #define vpx_highbd_filter_block1d8_v4_avg_sse2 … #define vpx_highbd_filter_block1d8_h4_avg_sse2 … #define vpx_highbd_filter_block1d4_v4_avg_sse2 … #define vpx_highbd_filter_block1d4_h4_avg_sse2 … // From vpx_dsp/x86/vpx_high_subpixel_bilinear_sse2.asm. highbd_filter8_1dfunction vpx_highbd_filter_block1d16_v2_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d16_h2_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d8_v2_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d8_h2_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d4_v2_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d4_h2_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d16_v2_avg_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d16_h2_avg_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d8_v2_avg_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d8_h2_avg_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d4_v2_avg_sse2; highbd_filter8_1dfunction vpx_highbd_filter_block1d4_h2_avg_sse2; // void vpx_highbd_convolve8_horiz_sse2(const uint8_t *src, // ptrdiff_t src_stride, // uint8_t *dst, // ptrdiff_t dst_stride, // const int16_t *filter_x, // int x_step_q4, // const int16_t *filter_y, // int y_step_q4, // int w, int h, int bd); // void vpx_highbd_convolve8_vert_sse2(const uint8_t *src, // ptrdiff_t src_stride, // uint8_t *dst, // ptrdiff_t dst_stride, // const int16_t *filter_x, // int x_step_q4, // const int16_t *filter_y, // int y_step_q4, // int w, int h, int bd); // void vpx_highbd_convolve8_avg_horiz_sse2(const uint8_t *src, // ptrdiff_t src_stride, // uint8_t *dst, // ptrdiff_t dst_stride, // const int16_t *filter_x, // int x_step_q4, // const int16_t *filter_y, // int y_step_q4, // int w, int h, int bd); // void vpx_highbd_convolve8_avg_vert_sse2(const uint8_t *src, // ptrdiff_t src_stride, // uint8_t *dst, // ptrdiff_t dst_stride, // const int16_t *filter_x, // int x_step_q4, // const int16_t *filter_y, // int y_step_q4, // int w, int h, int bd); HIGH_FUN_CONV_1D(…) HIGH_FUN_CONV_1D(…) HIGH_FUN_CONV_1D(…) HIGH_FUN_CONV_1D(…) // void vpx_highbd_convolve8_sse2(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, int bd); // void vpx_highbd_convolve8_avg_sse2(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, int bd); HIGH_FUN_CONV_2D(…) HIGH_FUN_CONV_2D(…) #endif // CONFIG_VP9_HIGHBITDEPTH && VPX_ARCH_X86_64
Codebase Browser
chromium