From d22a504d11a15dc3eab666859db0046b5a7d75c5 Mon Sep 17 00:00:00 2001 From: Scott LaVarnway Date: Wed, 11 Sep 2013 13:56:32 -0400 Subject: [PATCH] Improved 8t filters Reformatted version of a patch submitted by Erik/Tamar from Intel. For the test clips used, the decoder performance improved by ~2%. Change-Id: Ifbc37ac6311bca9ff1cfefe3f2e9b7f13a4a511b --- vp9/common/x86/vp9_asm_stubs.c | 94 +++-- vp9/common/x86/vp9_subpixel_8t_intrin_ssse3.c | 376 ++++++++++++++++++ vp9/vp9_common.mk | 1 + 3 files changed, 447 insertions(+), 24 deletions(-) create mode 100644 vp9/common/x86/vp9_subpixel_8t_intrin_ssse3.c diff --git a/vp9/common/x86/vp9_asm_stubs.c b/vp9/common/x86/vp9_asm_stubs.c index 3f1c19828..f9f00b937 100644 --- a/vp9/common/x86/vp9_asm_stubs.c +++ b/vp9/common/x86/vp9_asm_stubs.c @@ -13,6 +13,7 @@ #include "./vpx_config.h" #include "./vp9_rtcd.h" #include "vpx_ports/mem.h" + /////////////////////////////////////////////////////////////////////////// // the mmx function that does the bilinear filtering and var calculation // // int one pass // @@ -36,6 +37,8 @@ DECLARE_ALIGNED(16, const short, vp9_bilinear_filters_mmx[16][8]) = { { 8, 8, 8, 8, 120, 120, 120, 120 } }; + +int num_func_entry = 0; #if HAVE_SSSE3 void vp9_filter_block1d16_v8_ssse3(const unsigned char *src_ptr, const unsigned int src_pitch, @@ -44,6 +47,13 @@ void vp9_filter_block1d16_v8_ssse3(const unsigned char *src_ptr, unsigned int output_height, const short *filter); +void vp9_filter_block1d16_v8_intrin_ssse3(const unsigned char *src_ptr, + const unsigned int src_pitch, + unsigned char *output_ptr, + unsigned int out_pitch, + unsigned int output_height, + const short *filter); + void vp9_filter_block1d16_h8_ssse3(const unsigned char *src_ptr, const unsigned int src_pitch, unsigned char *output_ptr, @@ -51,6 +61,13 @@ void vp9_filter_block1d16_h8_ssse3(const unsigned char *src_ptr, unsigned int output_height, const short *filter); +void vp9_filter_block1d16_h8_intrin_ssse3(const unsigned char *src_ptr, + const unsigned int src_pitch, + unsigned char *output_ptr, + unsigned int out_pitch, + unsigned int output_height, + const short *filter); + void vp9_filter_block1d8_v8_ssse3(const unsigned char *src_ptr, const unsigned int src_pitch, unsigned char *output_ptr, @@ -58,6 +75,13 @@ void vp9_filter_block1d8_v8_ssse3(const unsigned char *src_ptr, unsigned int output_height, const short *filter); +void vp9_filter_block1d8_v8_intrin_ssse3(const unsigned char *src_ptr, + const unsigned int src_pitch, + unsigned char *output_ptr, + unsigned int out_pitch, + unsigned int output_height, + const short *filter); + void vp9_filter_block1d8_h8_ssse3(const unsigned char *src_ptr, const unsigned int src_pitch, unsigned char *output_ptr, @@ -65,6 +89,13 @@ void vp9_filter_block1d8_h8_ssse3(const unsigned char *src_ptr, unsigned int output_height, const short *filter); +void vp9_filter_block1d8_h8_intrin_ssse3(const unsigned char *src_ptr, + const unsigned int src_pitch, + unsigned char *output_ptr, + unsigned int out_pitch, + unsigned int output_height, + const short *filter); + void vp9_filter_block1d4_v8_ssse3(const unsigned char *src_ptr, const unsigned int src_pitch, unsigned char *output_ptr, @@ -72,6 +103,14 @@ void vp9_filter_block1d4_v8_ssse3(const unsigned char *src_ptr, unsigned int output_height, const short *filter); +void vp9_filter_block1d4_v8_intrin_ssse3(const unsigned char *src_ptr, + const unsigned int src_pitch, + unsigned char *output_ptr, + unsigned int out_pitch, + unsigned int output_height, + const short *filter); + + void vp9_filter_block1d4_h8_ssse3(const unsigned char *src_ptr, const unsigned int src_pitch, unsigned char *output_ptr, @@ -79,6 +118,13 @@ void vp9_filter_block1d4_h8_ssse3(const unsigned char *src_ptr, unsigned int output_height, const short *filter); +void vp9_filter_block1d4_h8_intrin_ssse3(const unsigned char *src_ptr, + const unsigned int src_pitch, + unsigned char *output_ptr, + unsigned int out_pitch, + unsigned int output_height, + const short *filter); + void vp9_filter_block1d16_v8_avg_ssse3(const unsigned char *src_ptr, const unsigned int src_pitch, unsigned char *output_ptr, @@ -129,7 +175,7 @@ void vp9_convolve8_horiz_ssse3(const uint8_t *src, ptrdiff_t src_stride, /* Ensure the filter can be compressed to int16_t. */ if (x_step_q4 == 16 && filter_x[3] != 128) { while (w >= 16) { - vp9_filter_block1d16_h8_ssse3(src, src_stride, + vp9_filter_block1d16_h8_intrin_ssse3(src, src_stride, dst, dst_stride, h, filter_x); src += 16; @@ -137,7 +183,7 @@ void vp9_convolve8_horiz_ssse3(const uint8_t *src, ptrdiff_t src_stride, w -= 16; } while (w >= 8) { - vp9_filter_block1d8_h8_ssse3(src, src_stride, + vp9_filter_block1d8_h8_intrin_ssse3(src, src_stride, dst, dst_stride, h, filter_x); src += 8; @@ -145,7 +191,7 @@ void vp9_convolve8_horiz_ssse3(const uint8_t *src, ptrdiff_t src_stride, w -= 8; } while (w >= 4) { - vp9_filter_block1d4_h8_ssse3(src, src_stride, + vp9_filter_block1d4_h8_intrin_ssse3(src, src_stride, dst, dst_stride, h, filter_x); src += 4; @@ -167,25 +213,25 @@ void vp9_convolve8_vert_ssse3(const uint8_t *src, ptrdiff_t src_stride, int w, int h) { if (y_step_q4 == 16 && filter_y[3] != 128) { while (w >= 16) { - vp9_filter_block1d16_v8_ssse3(src - src_stride * 3, src_stride, - dst, dst_stride, - h, filter_y); + vp9_filter_block1d16_v8_intrin_ssse3(src - src_stride * 3, src_stride, + dst, dst_stride, + h, filter_y); src += 16; dst += 16; w -= 16; } while (w >= 8) { - vp9_filter_block1d8_v8_ssse3(src - src_stride * 3, src_stride, - dst, dst_stride, - h, filter_y); + vp9_filter_block1d8_v8_intrin_ssse3(src - src_stride * 3, src_stride, + dst, dst_stride, + h, filter_y); src += 8; dst += 8; w -= 8; } while (w >= 4) { - vp9_filter_block1d4_v8_ssse3(src - src_stride * 3, src_stride, - dst, dst_stride, - h, filter_y); + vp9_filter_block1d4_v8_intrin_ssse3(src - src_stride * 3, src_stride, + dst, dst_stride, + h, filter_y); src += 4; dst += 4; w -= 4; @@ -206,24 +252,24 @@ void vp9_convolve8_avg_horiz_ssse3(const uint8_t *src, ptrdiff_t src_stride, if (x_step_q4 == 16 && filter_x[3] != 128) { while (w >= 16) { vp9_filter_block1d16_h8_avg_ssse3(src, src_stride, - dst, dst_stride, - h, filter_x); + dst, dst_stride, + h, filter_x); src += 16; dst += 16; w -= 16; } while (w >= 8) { vp9_filter_block1d8_h8_avg_ssse3(src, src_stride, - dst, dst_stride, - h, filter_x); + dst, dst_stride, + h, filter_x); src += 8; dst += 8; w -= 8; } while (w >= 4) { vp9_filter_block1d4_h8_avg_ssse3(src, src_stride, - dst, dst_stride, - h, filter_x); + dst, dst_stride, + h, filter_x); src += 4; dst += 4; w -= 4; @@ -244,24 +290,24 @@ void vp9_convolve8_avg_vert_ssse3(const uint8_t *src, ptrdiff_t src_stride, if (y_step_q4 == 16 && filter_y[3] != 128) { while (w >= 16) { vp9_filter_block1d16_v8_avg_ssse3(src - src_stride * 3, src_stride, - dst, dst_stride, - h, filter_y); + dst, dst_stride, + h, filter_y); src += 16; dst += 16; w -= 16; } while (w >= 8) { vp9_filter_block1d8_v8_avg_ssse3(src - src_stride * 3, src_stride, - dst, dst_stride, - h, filter_y); + dst, dst_stride, + h, filter_y); src += 8; dst += 8; w -= 8; } while (w >= 4) { vp9_filter_block1d4_v8_avg_ssse3(src - src_stride * 3, src_stride, - dst, dst_stride, - h, filter_y); + dst, dst_stride, + h, filter_y); src += 4; dst += 4; w -= 4; diff --git a/vp9/common/x86/vp9_subpixel_8t_intrin_ssse3.c b/vp9/common/x86/vp9_subpixel_8t_intrin_ssse3.c new file mode 100644 index 000000000..4242dab88 --- /dev/null +++ b/vp9/common/x86/vp9_subpixel_8t_intrin_ssse3.c @@ -0,0 +1,376 @@ +#include +#include +#include +#include "vpx_ports/mem.h" +#include "vpx_ports/emmintrin_compat.h" +#pragma GCC push_options +#pragma GCC optimize("unroll-loops") + +DECLARE_ALIGNED(16, const unsigned char, filt1_4_h8[16]) = { + 0, 1, 1, 2, 2, 3, 3, 4, 2, 3, 3, 4, 4, 5, 5, 6 +}; +DECLARE_ALIGNED(16, const unsigned char, filt2_4_h8[16]) = { + 4, 5, 5, 6, 6, 7, 7, 8, 6, 7, 7, 8, 8, 9, 9, 10 +}; +DECLARE_ALIGNED(16, const unsigned char, filt1_global[16]) = { + 0, 1, 1, 2, 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8 +}; +DECLARE_ALIGNED(16, const unsigned char, filt2_global[16]) = { + 2, 3, 3, 4, 4, 5, 5, 6, 6, 7, 7, 8, 8, 9, 9, 10 +}; +DECLARE_ALIGNED(16, const unsigned char, filt3_global[16]) = { + 4, 5, 5, 6, 6, 7, 7, 8, 8, 9, 9, 10, 10, 11, 11, 12 +}; +DECLARE_ALIGNED(16, const unsigned char, filt4_global[16]) = { + 6, 7, 7, 8, 8, 9, 9, 10, 10, 11, 11, 12, 12, 13, 13, 14 +}; + +void vp9_filter_block1d4_h8_intrin_ssse3(unsigned char *src_ptr, + unsigned int src_pixels_per_line, + unsigned char *output_ptr, + unsigned int output_pitch, + unsigned int output_height, + short *filter) { + __m128i addFilterReg127, filtersReg, firstFilters, secondFilters, + thirdFilters, forthFilters, srcReg, srcRegFilt1, srcRegFilt2, + srcRegFilt3, srcRegFilt4; + unsigned int i; + + addFilterReg127 = _mm_shuffle_epi32(_mm_cvtsi32_si128((int)0x0400040u), 0); + filtersReg = _mm_loadu_si128((__m128i *)filter); + filtersReg = _mm_packs_epi16(filtersReg, filtersReg); + firstFilters = _mm_shufflelo_epi16(filtersReg, 0); + secondFilters = _mm_shufflelo_epi16(filtersReg, 0xAAu); + firstFilters = _mm_shufflehi_epi16(firstFilters, 0x55u); + secondFilters = _mm_shufflehi_epi16(secondFilters, 0xFFu); + thirdFilters =_mm_load_si128((__m128i const *)filt1_4_h8); + forthFilters = _mm_load_si128((__m128i const *)filt2_4_h8); + + for (i = 0; i < output_height; i++) { + srcReg = _mm_loadu_si128((__m128i *)(src_ptr-3)); + srcRegFilt1= _mm_shuffle_epi8(srcReg, thirdFilters); + srcRegFilt2= _mm_shuffle_epi8(srcReg, forthFilters); + srcRegFilt1 = _mm_maddubs_epi16(srcRegFilt1, firstFilters); + srcRegFilt2 = _mm_maddubs_epi16(srcRegFilt2, secondFilters); + srcRegFilt3 = _mm_srli_si128(srcRegFilt1, 8); + srcRegFilt4 = _mm_srli_si128(srcRegFilt2, 8); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, srcRegFilt3); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, srcRegFilt4); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, srcRegFilt2); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, addFilterReg127); + srcRegFilt1 = _mm_srai_epi16(srcRegFilt1, 7); + srcRegFilt1 = _mm_packus_epi16(srcRegFilt1, srcRegFilt1); + src_ptr += src_pixels_per_line; + *((int*)&output_ptr[0])= _mm_cvtsi128_si32(srcRegFilt1); + output_ptr += output_pitch; + } +} + +void vp9_filter_block1d8_h8_intrin_ssse3(unsigned char *src_ptr, + unsigned int src_pixels_per_line, + unsigned char *output_ptr, + unsigned int output_pitch, + unsigned int output_height, + short *filter) { + __m128i addFilterReg127, filtersReg, firstFilters, secondFilters, + thirdFilters, forthFilters, srcReg, srcRegFilt1, srcRegFilt2, + srcRegFilt3, srcRegFilt4; + __m128i filt1Reg, filt2Reg, filt3Reg, filt4Reg; + unsigned int i; + + addFilterReg127 = _mm_shuffle_epi32(_mm_cvtsi32_si128((int)0x0400040u), 0); + filtersReg = _mm_loadu_si128((__m128i *)filter); + filtersReg =_mm_packs_epi16(filtersReg, filtersReg); + + firstFilters = _mm_shuffle_epi8(filtersReg, _mm_set1_epi16(0x100u)); + secondFilters = _mm_shuffle_epi8(filtersReg, _mm_set1_epi16(0x302u)); + thirdFilters = _mm_shuffle_epi8(filtersReg, _mm_set1_epi16(0x504u)); + forthFilters = _mm_shuffle_epi8(filtersReg, _mm_set1_epi16(0x706u)); + + filt1Reg = _mm_load_si128((__m128i const *)filt1_global); + filt2Reg = _mm_load_si128((__m128i const *)filt2_global); + filt3Reg = _mm_load_si128((__m128i const *)filt3_global); + filt4Reg = _mm_load_si128((__m128i const *)filt4_global); + + for (i = 0 ; i < output_height ; i++) { + srcReg = _mm_loadu_si128((__m128i *)(src_ptr-3)); + srcRegFilt1= _mm_shuffle_epi8(srcReg, filt1Reg); + srcRegFilt2= _mm_shuffle_epi8(srcReg, filt2Reg); + srcRegFilt1 = _mm_maddubs_epi16(srcRegFilt1, firstFilters); + srcRegFilt2 = _mm_maddubs_epi16(srcRegFilt2, secondFilters); + srcRegFilt3= _mm_shuffle_epi8(srcReg, filt3Reg); + srcRegFilt4= _mm_shuffle_epi8(srcReg, filt4Reg); + srcRegFilt3 = _mm_maddubs_epi16(srcRegFilt3, thirdFilters); + srcRegFilt4 = _mm_maddubs_epi16(srcRegFilt4, forthFilters); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, srcRegFilt2); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, srcRegFilt4); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, srcRegFilt3); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, addFilterReg127); + srcRegFilt1 = _mm_srai_epi16(srcRegFilt1, 7); + srcRegFilt1 = _mm_packus_epi16(srcRegFilt1, srcRegFilt1); + src_ptr += src_pixels_per_line; + _mm_storel_epi64((__m128i *)&output_ptr[0], srcRegFilt1); + output_ptr += output_pitch; + } +} + +void vp9_filter_block1d16_h8_intrin_ssse3(unsigned char *src_ptr, + unsigned int src_pixels_per_line, + unsigned char *output_ptr, + unsigned int output_pitch, + unsigned int output_height, + short *filter) { + __m128i addFilterReg127, filtersReg, firstFilters, secondFilters, + thirdFilters, forthFilters, srcReg1, srcReg2, srcRegFilt1_1, + srcRegFilt2_1, srcRegFilt2, srcRegFilt3; + __m128i filt1Reg, filt2Reg, filt3Reg, filt4Reg; + unsigned int i; + + addFilterReg127 = _mm_shuffle_epi32(_mm_cvtsi32_si128((int)0x0400040u), 0); + filtersReg = _mm_loadu_si128((__m128i *)filter); + filtersReg =_mm_packs_epi16(filtersReg, filtersReg); + + firstFilters = _mm_shuffle_epi8(filtersReg, _mm_set1_epi16(0x100u)); + secondFilters = _mm_shuffle_epi8(filtersReg, _mm_set1_epi16(0x302u)); + thirdFilters = _mm_shuffle_epi8(filtersReg, _mm_set1_epi16(0x504u)); + forthFilters = _mm_shuffle_epi8(filtersReg, _mm_set1_epi16(0x706u)); + + filt1Reg = _mm_load_si128((__m128i const *)filt1_global); + filt2Reg = _mm_load_si128((__m128i const *)filt2_global); + filt3Reg = _mm_load_si128((__m128i const *)filt3_global); + filt4Reg = _mm_load_si128((__m128i const *)filt4_global); + + for (i = 0 ; i < output_height ; i++) { + srcReg1 = _mm_loadu_si128((__m128i *)(src_ptr-3)); + srcRegFilt1_1 = _mm_shuffle_epi8(srcReg1, filt1Reg); + srcRegFilt2 = _mm_shuffle_epi8(srcReg1, filt2Reg); + + srcRegFilt1_1 = _mm_maddubs_epi16(srcRegFilt1_1, firstFilters); + srcRegFilt2 = _mm_maddubs_epi16(srcRegFilt2, secondFilters); + + srcRegFilt1_1 = _mm_adds_epi16(srcRegFilt1_1, srcRegFilt2); + + srcRegFilt3 = _mm_shuffle_epi8(srcReg1, filt4Reg); + srcRegFilt2 = _mm_shuffle_epi8(srcReg1, filt3Reg); + + srcRegFilt3 = _mm_maddubs_epi16(srcRegFilt3, forthFilters); + srcRegFilt2 = _mm_maddubs_epi16(srcRegFilt2, thirdFilters); + + srcRegFilt1_1 = _mm_adds_epi16(srcRegFilt1_1, srcRegFilt3); + + srcReg2 = _mm_loadu_si128((__m128i *)(src_ptr+5)); + + srcRegFilt1_1 = _mm_adds_epi16(srcRegFilt1_1, srcRegFilt2); + + srcRegFilt2_1 = _mm_shuffle_epi8(srcReg2, filt1Reg); + srcRegFilt2 = _mm_shuffle_epi8(srcReg2, filt2Reg); + + srcRegFilt2_1 = _mm_maddubs_epi16(srcRegFilt2_1, firstFilters); + srcRegFilt2 = _mm_maddubs_epi16(srcRegFilt2, secondFilters); + + srcRegFilt2_1 = _mm_adds_epi16(srcRegFilt2_1, srcRegFilt2); + + srcRegFilt3 = _mm_shuffle_epi8(srcReg2, filt4Reg); + srcRegFilt2 = _mm_shuffle_epi8(srcReg2, filt3Reg); + + srcRegFilt3 = _mm_maddubs_epi16(srcRegFilt3, forthFilters); + srcRegFilt2 = _mm_maddubs_epi16(srcRegFilt2, thirdFilters); + + srcRegFilt2_1 = _mm_adds_epi16(srcRegFilt2_1, srcRegFilt3); + srcRegFilt2_1 = _mm_adds_epi16(srcRegFilt2_1, srcRegFilt2); + srcRegFilt1_1 = _mm_adds_epi16(srcRegFilt1_1, addFilterReg127); + srcRegFilt2_1 = _mm_adds_epi16(srcRegFilt2_1, addFilterReg127); + srcRegFilt1_1 = _mm_srai_epi16(srcRegFilt1_1, 7); + srcRegFilt2_1 = _mm_srai_epi16(srcRegFilt2_1, 7); + srcRegFilt1_1 = _mm_packus_epi16(srcRegFilt1_1, srcRegFilt2_1); + + src_ptr += src_pixels_per_line; + _mm_store_si128((__m128i*)output_ptr, srcRegFilt1_1); + output_ptr += output_pitch; + } +} + +void vp9_filter_block1d4_v8_intrin_ssse3(unsigned char *src_ptr, + unsigned int src_pitch, + unsigned char *output_ptr, + unsigned int out_pitch, + unsigned int output_height, + short *filter) { + __m128i addFilterReg127, filtersReg, firstFilters, secondFilters, srcRegFilt1, + srcRegFilt2, srcRegFilt3, srcRegFilt4; + unsigned int i; + + addFilterReg127 = _mm_shuffle_epi32(_mm_cvtsi32_si128((int)0x0400040u), 0); + filtersReg = _mm_loadu_si128((__m128i *)filter); + filtersReg =_mm_packs_epi16(filtersReg, filtersReg); + firstFilters = _mm_shufflelo_epi16(filtersReg, 0); + firstFilters = _mm_shufflehi_epi16(firstFilters, 0x55u); + secondFilters = _mm_shufflelo_epi16(filtersReg, 0xAAu); + secondFilters = _mm_shufflehi_epi16(secondFilters, 0xFFu); + + for (i = 0 ; i < output_height ; i++) { + srcRegFilt1 = _mm_cvtsi32_si128(*((int*)&src_ptr[0])); + srcRegFilt2 = _mm_cvtsi32_si128(*((int*)&(src_ptr+src_pitch)[0])); + + srcRegFilt1 = _mm_unpacklo_epi8(srcRegFilt1, srcRegFilt2); + + srcRegFilt2 = _mm_cvtsi32_si128(*((int*)&(src_ptr+src_pitch*2)[0])); + srcRegFilt3 = _mm_cvtsi32_si128(*((int*)&(src_ptr+src_pitch*3)[0])); + + srcRegFilt2 = _mm_unpacklo_epi8(srcRegFilt2, srcRegFilt3); + + srcRegFilt3 = _mm_cvtsi32_si128(*((int*)&(src_ptr+src_pitch*4)[0])); + srcRegFilt4 = _mm_cvtsi32_si128(*((int*)&(src_ptr+src_pitch*5)[0])); + + srcRegFilt3 = _mm_unpacklo_epi8(srcRegFilt3, srcRegFilt4); + srcRegFilt1 = _mm_unpacklo_epi64(srcRegFilt1, srcRegFilt2); + + srcRegFilt4 = _mm_cvtsi32_si128(*((int*)&(src_ptr+src_pitch*6)[0])); + srcRegFilt2 = _mm_cvtsi32_si128(*((int*)&(src_ptr+src_pitch*7)[0])); + + srcRegFilt4 = _mm_unpacklo_epi8(srcRegFilt4, srcRegFilt2); + srcRegFilt3 = _mm_unpacklo_epi64(srcRegFilt3, srcRegFilt4); + + srcRegFilt1 = _mm_maddubs_epi16(srcRegFilt1, firstFilters); + srcRegFilt3 = _mm_maddubs_epi16(srcRegFilt3, secondFilters); + srcRegFilt2 = _mm_srli_si128(srcRegFilt1, 8); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, _mm_srli_si128(srcRegFilt3, 8)); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, srcRegFilt2); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, srcRegFilt3); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, addFilterReg127); + srcRegFilt1 = _mm_srai_epi16(srcRegFilt1, 7); + srcRegFilt1 = _mm_packus_epi16(srcRegFilt1, srcRegFilt1); + + src_ptr += src_pitch; + *((int*)&output_ptr[0])= _mm_cvtsi128_si32(srcRegFilt1); + output_ptr += out_pitch; + } +} + +void vp9_filter_block1d8_v8_intrin_ssse3(unsigned char *src_ptr, + unsigned int src_pitch, + unsigned char *output_ptr, + unsigned int out_pitch, + unsigned int output_height, + short *filter) { + __m128i addFilterReg127, filtersReg, firstFilters, secondFilters, + thirdFilters, forthFilters, srcRegFilt1, srcRegFilt2, srcRegFilt3, + srcRegFilt4, srcRegFilt5, srcRegFilt6; + unsigned int i; + + addFilterReg127 = _mm_shuffle_epi32(_mm_cvtsi32_si128((int)0x0400040u), 0); + filtersReg = _mm_loadu_si128((__m128i *)filter); + filtersReg =_mm_packs_epi16(filtersReg, filtersReg); + + firstFilters = _mm_shuffle_epi8(filtersReg, _mm_set1_epi16(0x100u)); + secondFilters = _mm_shuffle_epi8(filtersReg, _mm_set1_epi16(0x302u)); + thirdFilters = _mm_shuffle_epi8(filtersReg, _mm_set1_epi16(0x504u)); + forthFilters = _mm_shuffle_epi8(filtersReg, _mm_set1_epi16(0x706u)); + + for (i = 0 ; i < output_height ; i++) { + srcRegFilt1 = _mm_loadl_epi64((__m128i*)&src_ptr[0]); + srcRegFilt2 = _mm_loadl_epi64((__m128i*)&(src_ptr+src_pitch)[0]); + + srcRegFilt3 = _mm_loadl_epi64((__m128i*)&(src_ptr+src_pitch*2)[0]); + srcRegFilt4 = _mm_loadl_epi64((__m128i*)&(src_ptr+src_pitch*3)[0]); + + srcRegFilt1 = _mm_unpacklo_epi8(srcRegFilt1, srcRegFilt2); + + srcRegFilt3 = _mm_unpacklo_epi8(srcRegFilt3, srcRegFilt4); + + srcRegFilt2 = _mm_loadl_epi64((__m128i*)&(src_ptr+src_pitch*4)[0]); + srcRegFilt4 = _mm_loadl_epi64((__m128i*)&(src_ptr+src_pitch*5)[0]); + + srcRegFilt5 = _mm_loadl_epi64((__m128i*)&(src_ptr+src_pitch*6)[0]); + srcRegFilt6 = _mm_loadl_epi64((__m128i*)&(src_ptr+src_pitch*7)[0]); + + srcRegFilt2 = _mm_unpacklo_epi8(srcRegFilt2, srcRegFilt4); + srcRegFilt5 = _mm_unpacklo_epi8(srcRegFilt5, srcRegFilt6); + srcRegFilt1 = _mm_maddubs_epi16(srcRegFilt1, firstFilters); + srcRegFilt3 = _mm_maddubs_epi16(srcRegFilt3, secondFilters); + srcRegFilt2 = _mm_maddubs_epi16(srcRegFilt2, thirdFilters); + srcRegFilt5 = _mm_maddubs_epi16(srcRegFilt5, forthFilters); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, srcRegFilt5); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, srcRegFilt3); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, srcRegFilt2); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, addFilterReg127); + srcRegFilt1 = _mm_srai_epi16(srcRegFilt1, 7); + srcRegFilt1 = _mm_packus_epi16(srcRegFilt1, srcRegFilt1); + + src_ptr += src_pitch; + _mm_storel_epi64((__m128i *)&output_ptr[0], srcRegFilt1); + output_ptr += out_pitch; + } +} + +void vp9_filter_block1d16_v8_intrin_ssse3(unsigned char *src_ptr, + unsigned int src_pitch, + unsigned char *output_ptr, + unsigned int out_pitch, + unsigned int output_height, + short *filter) { + __m128i addFilterReg127, filtersReg, firstFilters, secondFilters, + thirdFilters, forthFilters, srcRegFilt1, srcRegFilt2, srcRegFilt3, + srcRegFilt4, srcRegFilt5, srcRegFilt6; + unsigned int i; + + addFilterReg127 = _mm_shuffle_epi32(_mm_cvtsi32_si128((int)0x0400040u), 0); + filtersReg = _mm_loadu_si128((__m128i *)filter); + filtersReg =_mm_packs_epi16(filtersReg, filtersReg); + + firstFilters = _mm_shuffle_epi8(filtersReg, _mm_set1_epi16(0x100u)); + secondFilters = _mm_shuffle_epi8(filtersReg, _mm_set1_epi16(0x302u)); + thirdFilters = _mm_shuffle_epi8(filtersReg, _mm_set1_epi16(0x504u)); + forthFilters = _mm_shuffle_epi8(filtersReg, _mm_set1_epi16(0x706u)); + + for (i = 0 ; i < output_height ; i++) { + srcRegFilt1 = _mm_loadu_si128((__m128i *)(src_ptr)); + srcRegFilt2 = _mm_loadu_si128((__m128i *)(src_ptr+src_pitch)); + + srcRegFilt3 = _mm_loadu_si128((__m128i *)(src_ptr+src_pitch*6)); + srcRegFilt4 = _mm_loadu_si128((__m128i *)(src_ptr+src_pitch*7)); + + srcRegFilt5 = _mm_unpacklo_epi8(srcRegFilt1, srcRegFilt2); + srcRegFilt6 = _mm_unpacklo_epi8(srcRegFilt3, srcRegFilt4); + srcRegFilt1 = _mm_unpackhi_epi8(srcRegFilt1, srcRegFilt2); + srcRegFilt3 = _mm_unpackhi_epi8(srcRegFilt3, srcRegFilt4); + + srcRegFilt5 = _mm_maddubs_epi16(srcRegFilt5, firstFilters); + srcRegFilt6 = _mm_maddubs_epi16(srcRegFilt6, forthFilters); + srcRegFilt1 = _mm_maddubs_epi16(srcRegFilt1, firstFilters); + srcRegFilt3 = _mm_maddubs_epi16(srcRegFilt3, forthFilters); + + srcRegFilt5 = _mm_adds_epi16(srcRegFilt5, srcRegFilt6); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, srcRegFilt3); + + srcRegFilt2 = _mm_loadu_si128((__m128i *)(src_ptr+src_pitch*2)); + srcRegFilt3 = _mm_loadu_si128((__m128i *)(src_ptr+src_pitch*3)); + + srcRegFilt4 = _mm_unpacklo_epi8(srcRegFilt2, srcRegFilt3); + srcRegFilt6 = _mm_unpackhi_epi8(srcRegFilt2, srcRegFilt3); + srcRegFilt4 = _mm_maddubs_epi16(srcRegFilt4, secondFilters); + srcRegFilt6 = _mm_maddubs_epi16(srcRegFilt6, secondFilters); + + srcRegFilt5 = _mm_adds_epi16(srcRegFilt5, srcRegFilt4); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, srcRegFilt6); + + srcRegFilt2 = _mm_loadu_si128((__m128i *)(src_ptr+src_pitch*4)); + srcRegFilt3 = _mm_loadu_si128((__m128i *)(src_ptr+src_pitch*5)); + + srcRegFilt4 = _mm_unpacklo_epi8(srcRegFilt2, srcRegFilt3); + srcRegFilt6 = _mm_unpackhi_epi8(srcRegFilt2, srcRegFilt3); + srcRegFilt4 = _mm_maddubs_epi16(srcRegFilt4, thirdFilters); + srcRegFilt6 = _mm_maddubs_epi16(srcRegFilt6, thirdFilters); + srcRegFilt5 = _mm_adds_epi16(srcRegFilt5, srcRegFilt4); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, srcRegFilt6); + srcRegFilt5 = _mm_adds_epi16(srcRegFilt5, addFilterReg127); + srcRegFilt1 = _mm_adds_epi16(srcRegFilt1, addFilterReg127); + srcRegFilt5 = _mm_srai_epi16(srcRegFilt5, 7); + srcRegFilt1 = _mm_srai_epi16(srcRegFilt1, 7); + srcRegFilt1 = _mm_packus_epi16(srcRegFilt5, srcRegFilt1); + + src_ptr += src_pitch; + _mm_store_si128((__m128i*)output_ptr, srcRegFilt1); + output_ptr += out_pitch; + } +} diff --git a/vp9/vp9_common.mk b/vp9/vp9_common.mk index c6daecca0..866d397e3 100644 --- a/vp9/vp9_common.mk +++ b/vp9/vp9_common.mk @@ -77,6 +77,7 @@ VP9_COMMON_SRCS-$(CONFIG_VP9_POSTPROC) += common/vp9_postproc.h VP9_COMMON_SRCS-$(CONFIG_VP9_POSTPROC) += common/vp9_postproc.c VP9_COMMON_SRCS-$(HAVE_MMX) += common/x86/vp9_loopfilter_mmx.asm VP9_COMMON_SRCS-$(HAVE_SSSE3) += common/x86/vp9_subpixel_8t_ssse3.asm +VP9_COMMON_SRCS-$(HAVE_SSSE3) += common/x86/vp9_subpixel_8t_intrin_ssse3.c ifeq ($(CONFIG_VP9_POSTPROC),yes) VP9_COMMON_SRCS-$(HAVE_MMX) += common/x86/vp9_postproc_mmx.asm VP9_COMMON_SRCS-$(HAVE_SSE2) += common/x86/vp9_postproc_sse2.asm -- 2.40.0