vpx/vp9/common/x86/vp9_asm_stubs.c
levytamar82 3068d7d944 SSSE3 convolution optimization
Optimizing all SSSE3 assembly for convolution:
1. vp9_filter_block1d4_h8_sse2
2. vp9_filter_block1d8_h8_sse2
3. vp9_filter_block1d16_h8_sse2
4. vp9_filter_block1d4_v8_sse2
5. vp9_filter_block1d8_v8_sse2
6. vp9_filter_block1d16_v8_sse2
my optimization include:
-processing 2x8 elements in one 128 bit register instead of processing
8 elements in one 128 bit register.
-removing unecessary loads.
This optimization gives between 2.4% user level gain for 480p input
and 1.6% user level gain for 720p.
This Optimization is done only for 64 bit

Change-Id: Ic07fce2f9360329b4f2d956efda1480ae958766b
2014-02-14 15:08:42 -07:00

340 lines
16 KiB
C

/*
* Copyright (c) 2014 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 <assert.h>
#include "./vpx_config.h"
#include "./vp9_rtcd.h"
#include "vpx_ports/mem.h"
typedef void filter8_1dfunction (
const unsigned char *src_ptr,
const ptrdiff_t src_pitch,
unsigned char *output_ptr,
ptrdiff_t out_pitch,
unsigned int output_height,
const short *filter
);
#define FUN_CONV_1D(name, step_q4, filter, dir, src_start, avg, opt) \
void vp9_convolve8_##name##_##opt(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) { \
if (step_q4 == 16 && filter[3] != 128) { \
if (filter[0] || filter[1] || filter[2]) { \
while (w >= 16) { \
vp9_filter_block1d16_##dir##8_##avg##opt(src_start, \
src_stride, \
dst, \
dst_stride, \
h, \
filter); \
src += 16; \
dst += 16; \
w -= 16; \
} \
while (w >= 8) { \
vp9_filter_block1d8_##dir##8_##avg##opt(src_start, \
src_stride, \
dst, \
dst_stride, \
h, \
filter); \
src += 8; \
dst += 8; \
w -= 8; \
} \
while (w >= 4) { \
vp9_filter_block1d4_##dir##8_##avg##opt(src_start, \
src_stride, \
dst, \
dst_stride, \
h, \
filter); \
src += 4; \
dst += 4; \
w -= 4; \
} \
} else { \
while (w >= 16) { \
vp9_filter_block1d16_##dir##2_##avg##opt(src, \
src_stride, \
dst, \
dst_stride, \
h, \
filter); \
src += 16; \
dst += 16; \
w -= 16; \
} \
while (w >= 8) { \
vp9_filter_block1d8_##dir##2_##avg##opt(src, \
src_stride, \
dst, \
dst_stride, \
h, \
filter); \
src += 8; \
dst += 8; \
w -= 8; \
} \
while (w >= 4) { \
vp9_filter_block1d4_##dir##2_##avg##opt(src, \
src_stride, \
dst, \
dst_stride, \
h, \
filter); \
src += 4; \
dst += 4; \
w -= 4; \
} \
} \
} \
if (w) { \
vp9_convolve8_##name##_c(src, src_stride, dst, dst_stride, \
filter_x, x_step_q4, filter_y, y_step_q4, \
w, h); \
} \
}
#define FUN_CONV_2D(avg, opt) \
void vp9_convolve8_##avg##opt(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) { \
assert(w <= 64); \
assert(h <= 64); \
if (x_step_q4 == 16 && y_step_q4 == 16) { \
if (filter_x[0] || filter_x[1] || filter_x[2] || filter_x[3] == 128 || \
filter_y[0] || filter_y[1] || filter_y[2] || filter_y[3] == 128) { \
DECLARE_ALIGNED_ARRAY(16, unsigned char, fdata2, 64 * 71); \
vp9_convolve8_horiz_##opt(src - 3 * src_stride, src_stride, fdata2, 64, \
filter_x, x_step_q4, filter_y, y_step_q4, \
w, h + 7); \
vp9_convolve8_##avg##vert_##opt(fdata2 + 3 * 64, 64, dst, dst_stride, \
filter_x, x_step_q4, filter_y, \
y_step_q4, w, h); \
} else { \
DECLARE_ALIGNED_ARRAY(16, unsigned char, fdata2, 64 * 65); \
vp9_convolve8_horiz_##opt(src, src_stride, fdata2, 64, \
filter_x, x_step_q4, filter_y, y_step_q4, \
w, h + 1); \
vp9_convolve8_##avg##vert_##opt(fdata2, 64, dst, dst_stride, \
filter_x, x_step_q4, filter_y, \
y_step_q4, w, h); \
} \
} else { \
vp9_convolve8_##avg##c(src, src_stride, dst, dst_stride, \
filter_x, x_step_q4, filter_y, y_step_q4, w, h); \
} \
}
#if HAVE_AVX2
filter8_1dfunction vp9_filter_block1d16_v8_avx2;
filter8_1dfunction vp9_filter_block1d16_h8_avx2;
filter8_1dfunction vp9_filter_block1d4_v8_ssse3;
#if (ARCH_X86_64)
filter8_1dfunction vp9_filter_block1d8_v8_intrin_ssse3;
filter8_1dfunction vp9_filter_block1d8_h8_intrin_ssse3;
filter8_1dfunction vp9_filter_block1d4_h8_intrin_ssse3;
#define vp9_filter_block1d8_v8_avx2 vp9_filter_block1d8_v8_intrin_ssse3
#define vp9_filter_block1d8_h8_avx2 vp9_filter_block1d8_h8_intrin_ssse3
#define vp9_filter_block1d4_h8_avx2 vp9_filter_block1d4_h8_intrin_ssse3
#else
filter8_1dfunction vp9_filter_block1d8_v8_ssse3;
filter8_1dfunction vp9_filter_block1d8_h8_ssse3;
filter8_1dfunction vp9_filter_block1d4_h8_ssse3;
#define vp9_filter_block1d8_v8_avx2 vp9_filter_block1d8_v8_ssse3
#define vp9_filter_block1d8_h8_avx2 vp9_filter_block1d8_h8_ssse3
#define vp9_filter_block1d4_h8_avx2 vp9_filter_block1d4_h8_ssse3
#endif
filter8_1dfunction vp9_filter_block1d16_v2_ssse3;
filter8_1dfunction vp9_filter_block1d16_h2_ssse3;
filter8_1dfunction vp9_filter_block1d8_v2_ssse3;
filter8_1dfunction vp9_filter_block1d8_h2_ssse3;
filter8_1dfunction vp9_filter_block1d4_v2_ssse3;
filter8_1dfunction vp9_filter_block1d4_h2_ssse3;
#define vp9_filter_block1d4_v8_avx2 vp9_filter_block1d4_v8_ssse3
#define vp9_filter_block1d16_v2_avx2 vp9_filter_block1d16_v2_ssse3
#define vp9_filter_block1d16_h2_avx2 vp9_filter_block1d16_h2_ssse3
#define vp9_filter_block1d8_v2_avx2 vp9_filter_block1d8_v2_ssse3
#define vp9_filter_block1d8_h2_avx2 vp9_filter_block1d8_h2_ssse3
#define vp9_filter_block1d4_v2_avx2 vp9_filter_block1d4_v2_ssse3
#define vp9_filter_block1d4_h2_avx2 vp9_filter_block1d4_h2_ssse3
// void vp9_convolve8_horiz_avx2(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);
// void vp9_convolve8_vert_avx2(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);
FUN_CONV_1D(horiz, x_step_q4, filter_x, h, src, , avx2);
FUN_CONV_1D(vert, y_step_q4, filter_y, v, src - src_stride * 3, , avx2);
// void vp9_convolve8_avx2(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);
FUN_CONV_2D(, avx2);
#endif
#if HAVE_SSSE3
#if (ARCH_X86_64)
filter8_1dfunction vp9_filter_block1d16_v8_intrin_ssse3;
filter8_1dfunction vp9_filter_block1d16_h8_intrin_ssse3;
filter8_1dfunction vp9_filter_block1d8_v8_intrin_ssse3;
filter8_1dfunction vp9_filter_block1d8_h8_intrin_ssse3;
filter8_1dfunction vp9_filter_block1d4_v8_ssse3;
filter8_1dfunction vp9_filter_block1d4_h8_intrin_ssse3;
#define vp9_filter_block1d16_v8_ssse3 vp9_filter_block1d16_v8_intrin_ssse3
#define vp9_filter_block1d16_h8_ssse3 vp9_filter_block1d16_h8_intrin_ssse3
#define vp9_filter_block1d8_v8_ssse3 vp9_filter_block1d8_v8_intrin_ssse3
#define vp9_filter_block1d8_h8_ssse3 vp9_filter_block1d8_h8_intrin_ssse3
#define vp9_filter_block1d4_h8_ssse3 vp9_filter_block1d4_h8_intrin_ssse3
#else
filter8_1dfunction vp9_filter_block1d16_v8_ssse3;
filter8_1dfunction vp9_filter_block1d16_h8_ssse3;
filter8_1dfunction vp9_filter_block1d8_v8_ssse3;
filter8_1dfunction vp9_filter_block1d8_h8_ssse3;
filter8_1dfunction vp9_filter_block1d4_v8_ssse3;
filter8_1dfunction vp9_filter_block1d4_h8_ssse3;
#endif
filter8_1dfunction vp9_filter_block1d16_v8_avg_ssse3;
filter8_1dfunction vp9_filter_block1d16_h8_avg_ssse3;
filter8_1dfunction vp9_filter_block1d8_v8_avg_ssse3;
filter8_1dfunction vp9_filter_block1d8_h8_avg_ssse3;
filter8_1dfunction vp9_filter_block1d4_v8_avg_ssse3;
filter8_1dfunction vp9_filter_block1d4_h8_avg_ssse3;
filter8_1dfunction vp9_filter_block1d16_v2_ssse3;
filter8_1dfunction vp9_filter_block1d16_h2_ssse3;
filter8_1dfunction vp9_filter_block1d8_v2_ssse3;
filter8_1dfunction vp9_filter_block1d8_h2_ssse3;
filter8_1dfunction vp9_filter_block1d4_v2_ssse3;
filter8_1dfunction vp9_filter_block1d4_h2_ssse3;
filter8_1dfunction vp9_filter_block1d16_v2_avg_ssse3;
filter8_1dfunction vp9_filter_block1d16_h2_avg_ssse3;
filter8_1dfunction vp9_filter_block1d8_v2_avg_ssse3;
filter8_1dfunction vp9_filter_block1d8_h2_avg_ssse3;
filter8_1dfunction vp9_filter_block1d4_v2_avg_ssse3;
filter8_1dfunction vp9_filter_block1d4_h2_avg_ssse3;
// void vp9_convolve8_horiz_ssse3(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);
// void vp9_convolve8_vert_ssse3(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);
// void vp9_convolve8_avg_horiz_ssse3(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);
// void vp9_convolve8_avg_vert_ssse3(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);
FUN_CONV_1D(horiz, x_step_q4, filter_x, h, src, , ssse3);
FUN_CONV_1D(vert, y_step_q4, filter_y, v, src - src_stride * 3, , ssse3);
FUN_CONV_1D(avg_horiz, x_step_q4, filter_x, h, src, avg_, ssse3);
FUN_CONV_1D(avg_vert, y_step_q4, filter_y, v, src - src_stride * 3, avg_,
ssse3);
// void vp9_convolve8_ssse3(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);
// void vp9_convolve8_avg_ssse3(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);
FUN_CONV_2D(, ssse3);
FUN_CONV_2D(avg_ , ssse3);
#endif
#if HAVE_SSE2
filter8_1dfunction vp9_filter_block1d16_v8_sse2;
filter8_1dfunction vp9_filter_block1d16_h8_sse2;
filter8_1dfunction vp9_filter_block1d8_v8_sse2;
filter8_1dfunction vp9_filter_block1d8_h8_sse2;
filter8_1dfunction vp9_filter_block1d4_v8_sse2;
filter8_1dfunction vp9_filter_block1d4_h8_sse2;
filter8_1dfunction vp9_filter_block1d16_v8_avg_sse2;
filter8_1dfunction vp9_filter_block1d16_h8_avg_sse2;
filter8_1dfunction vp9_filter_block1d8_v8_avg_sse2;
filter8_1dfunction vp9_filter_block1d8_h8_avg_sse2;
filter8_1dfunction vp9_filter_block1d4_v8_avg_sse2;
filter8_1dfunction vp9_filter_block1d4_h8_avg_sse2;
filter8_1dfunction vp9_filter_block1d16_v2_sse2;
filter8_1dfunction vp9_filter_block1d16_h2_sse2;
filter8_1dfunction vp9_filter_block1d8_v2_sse2;
filter8_1dfunction vp9_filter_block1d8_h2_sse2;
filter8_1dfunction vp9_filter_block1d4_v2_sse2;
filter8_1dfunction vp9_filter_block1d4_h2_sse2;
filter8_1dfunction vp9_filter_block1d16_v2_avg_sse2;
filter8_1dfunction vp9_filter_block1d16_h2_avg_sse2;
filter8_1dfunction vp9_filter_block1d8_v2_avg_sse2;
filter8_1dfunction vp9_filter_block1d8_h2_avg_sse2;
filter8_1dfunction vp9_filter_block1d4_v2_avg_sse2;
filter8_1dfunction vp9_filter_block1d4_h2_avg_sse2;
// void vp9_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);
// void vp9_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);
// void vp9_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);
// void vp9_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);
FUN_CONV_1D(horiz, x_step_q4, filter_x, h, src, , sse2);
FUN_CONV_1D(vert, y_step_q4, filter_y, v, src - src_stride * 3, , sse2);
FUN_CONV_1D(avg_horiz, x_step_q4, filter_x, h, src, avg_, sse2);
FUN_CONV_1D(avg_vert, y_step_q4, filter_y, v, src - src_stride * 3, avg_, sse2);
// void vp9_convolve8_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);
// void vp9_convolve8_avg_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);
FUN_CONV_2D(, sse2);
FUN_CONV_2D(avg_ , sse2);
#endif