simd-team/speccpulib
7
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

speccpu/x264: add x264 mc src, add printf to weight->i_scale

Browse Source
This commit is contained in:
Yingchi Long
2025-04-14 23:40:55 +08:00
parent 133b4dfae4
commit 710122cd1a
3 changed files with 609 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

7
pyproject.toml
View File

@@ -5,3 +5,10 @@ description = "Scripts that helps debugging SPEC CPU benchmark suite."
readme = "README.md"
requires-python = ">=3.9"
dependencies = []
[tool.setuptools]
packages = { find = { where = ["src"] } }
include-package-data = true
[tool.setuptools.package-data]
"speccpu" = ["*.c"]

547
src/speccpu/mc.c Normal file
View File

@@ -0,0 +1,547 @@
/*****************************************************************************
* mc.c: h264 encoder library (Motion Compensation)
*****************************************************************************
* Copyright (C) 2003-2008 x264 project
*
* Authors: Laurent Aimar <fenrir@via.ecp.fr>
* Loren Merritt <lorenm@u.washington.edu>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111, USA.
*****************************************************************************/
#include "common.h"
#ifdef HAVE_MMX
#include "x86/mc.h"
#endif
#ifdef ARCH_PPC
#include "ppc/mc.h"
#endif
#ifdef ARCH_ARM
#include "arm/mc.h"
#endif
static inline void pixel_avg( uint8_t *dst, int i_dst_stride,
uint8_t *src1, int i_src1_stride,
uint8_t *src2, int i_src2_stride,
int i_width, int i_height )
{
for( int y = 0; y < i_height; y++ )
{
for( int x = 0; x < i_width; x++ )
dst[x] = ( src1[x] + src2[x] + 1 ) >> 1;
dst += i_dst_stride;
src1 += i_src1_stride;
src2 += i_src2_stride;
}
}
static inline void pixel_avg_wxh( uint8_t *dst, int i_dst, uint8_t *src1, int i_src1, uint8_t *src2, int i_src2, int width, int height )
{
for( int y = 0; y < height; y++ )
{
for( int x = 0; x < width; x++ )
dst[x] = ( src1[x] + src2[x] + 1 ) >> 1;
src1 += i_src1;
src2 += i_src2;
dst += i_dst;
}
}
/* Implicit weighted bipred only:
* assumes log2_denom = 5, offset = 0, weight1 + weight2 = 64 */
#define op_scale2(x) dst[x] = x264_clip_uint8( (src1[x]*i_weight1 + src2[x]*i_weight2 + (1<<5)) >> 6 )
static inline void pixel_avg_weight_wxh( uint8_t *dst, int i_dst, uint8_t *src1, int i_src1, uint8_t *src2, int i_src2, int width, int height, int i_weight1 )
{
const int i_weight2 = 64 - i_weight1;
for( int y = 0; y<height; y++, dst += i_dst, src1 += i_src1, src2 += i_src2 )
{
op_scale2(0);
op_scale2(1);
if(width==2) continue;
op_scale2(2);
op_scale2(3);
if(width==4) continue;
op_scale2(4);
op_scale2(5);
op_scale2(6);
op_scale2(7);
if(width==8) continue;
op_scale2(8);
op_scale2(9);
op_scale2(10);
op_scale2(11);
op_scale2(12);
op_scale2(13);
op_scale2(14);
op_scale2(15);
}
}
#undef op_scale2
#define PIXEL_AVG_C( name, width, height ) \
static void name( uint8_t *pix1, int i_stride_pix1, \
uint8_t *pix2, int i_stride_pix2, \
uint8_t *pix3, int i_stride_pix3, int weight ) \
{ \
if( weight == 32 ) \
pixel_avg_wxh( pix1, i_stride_pix1, pix2, i_stride_pix2, pix3, i_stride_pix3, width, height ); \
else \
pixel_avg_weight_wxh( pix1, i_stride_pix1, pix2, i_stride_pix2, pix3, i_stride_pix3, width, height, weight ); \
}
PIXEL_AVG_C( pixel_avg_16x16, 16, 16 )
PIXEL_AVG_C( pixel_avg_16x8, 16, 8 )
PIXEL_AVG_C( pixel_avg_8x16, 8, 16 )
PIXEL_AVG_C( pixel_avg_8x8, 8, 8 )
PIXEL_AVG_C( pixel_avg_8x4, 8, 4 )
PIXEL_AVG_C( pixel_avg_4x8, 4, 8 )
PIXEL_AVG_C( pixel_avg_4x4, 4, 4 )
PIXEL_AVG_C( pixel_avg_4x2, 4, 2 )
PIXEL_AVG_C( pixel_avg_2x4, 2, 4 )
PIXEL_AVG_C( pixel_avg_2x2, 2, 2 )
static void x264_weight_cache( x264_t *h, x264_weight_t *w )
{
w->weightfn = h->mc.weight;
}
#define opscale(x) dst[x] = x264_clip_uint8( ((src[x] * weight->i_scale + (1<<(weight->i_denom - 1))) >> weight->i_denom) + weight->i_offset )
#define opscale_noden(x) dst[x] = x264_clip_uint8( src[x] * weight->i_scale + weight->i_offset )
static inline void mc_weight( uint8_t *dst, int i_dst_stride, uint8_t *src, int i_src_stride, const x264_weight_t *weight, int i_width, int i_height )
{
if( weight->i_denom >= 1 )
{
for( int y = 0; y < i_height; y++, dst += i_dst_stride, src += i_src_stride )
for( int x = 0; x < i_width; x++ )
opscale( x );
}
else
{
for( int y = 0; y < i_height; y++, dst += i_dst_stride, src += i_src_stride )
for( int x = 0; x < i_width; x++ )
opscale_noden( x );
}
}
#define MC_WEIGHT_C( name, lx ) \
static void name( uint8_t *dst, int i_dst_stride, uint8_t *src, int i_src_stride, const x264_weight_t *weight, int height ) \
{ \
if( weight->i_denom >= 1 ) \
{ \
for( int y = 0; y < height; y++, dst += i_dst_stride, src += i_src_stride ) \
for( int x = 0; x < lx; x++ ) \
opscale( x ); \
} \
else \
{ \
for( int y = 0; y < height; y++, dst += i_dst_stride, src += i_src_stride ) \
for( int x = 0; x < lx; x++ ) \
opscale_noden( x ); \
} \
}
MC_WEIGHT_C( mc_weight_w20, 20 )
MC_WEIGHT_C( mc_weight_w16, 16 )
MC_WEIGHT_C( mc_weight_w12, 12 )
MC_WEIGHT_C( mc_weight_w8, 8 )
MC_WEIGHT_C( mc_weight_w4, 4 )
MC_WEIGHT_C( mc_weight_w2, 2 )
static weight_fn_t x264_mc_weight_wtab[6] =
{
mc_weight_w2,
mc_weight_w4,
mc_weight_w8,
mc_weight_w12,
mc_weight_w16,
mc_weight_w20,
};
const x264_weight_t weight_none[3] = { {{0,},{0,},0,0,0,0} };
static void mc_copy( uint8_t *src, int i_src_stride, uint8_t *dst, int i_dst_stride, int i_width, int i_height )
{
for( int y = 0; y < i_height; y++ )
{
memcpy( dst, src, i_width );
src += i_src_stride;
dst += i_dst_stride;
}
}
#define TAPFILTER(pix, d) ((pix)[x-2*d] + (pix)[x+3*d] - 5*((pix)[x-d] + (pix)[x+2*d]) + 20*((pix)[x] + (pix)[x+d]))
static void hpel_filter( uint8_t *dsth, uint8_t *dstv, uint8_t *dstc, uint8_t *src,
int stride, int width, int height, int16_t *buf )
{
for( int y = 0; y < height; y++ )
{
for( int x = -2; x < width+3; x++ )
{
int v = TAPFILTER(src,stride);
dstv[x] = x264_clip_uint8( (v + 16) >> 5 );
buf[x+2] = v;
}
for( int x = 0; x < width; x++ )
dstc[x] = x264_clip_uint8( (TAPFILTER(buf+2,1) + 512) >> 10 );
for( int x = 0; x < width; x++ )
dsth[x] = x264_clip_uint8( (TAPFILTER(src,1) + 16) >> 5 );
dsth += stride;
dstv += stride;
dstc += stride;
src += stride;
}
}
static const uint8_t hpel_ref0[16] = {0,1,1,1,0,1,1,1,2,3,3,3,0,1,1,1};
static const uint8_t hpel_ref1[16] = {0,0,0,0,2,2,3,2,2,2,3,2,2,2,3,2};
static void mc_luma( uint8_t *dst, int i_dst_stride,
uint8_t *src[4], int i_src_stride,
int mvx, int mvy,
int i_width, int i_height, const x264_weight_t *weight )
{
int qpel_idx = ((mvy&3)<<2) + (mvx&3);
int offset = (mvy>>2)*i_src_stride + (mvx>>2);
uint8_t *src1 = src[hpel_ref0[qpel_idx]] + offset + ((mvy&3) == 3) * i_src_stride;
if( qpel_idx & 5 ) /* qpel interpolation needed */
{
uint8_t *src2 = src[hpel_ref1[qpel_idx]] + offset + ((mvx&3) == 3);
pixel_avg( dst, i_dst_stride, src1, i_src_stride,
src2, i_src_stride, i_width, i_height );
if( weight->weightfn )
mc_weight( dst, i_dst_stride, dst, i_dst_stride, weight, i_width, i_height );
}
else if( weight->weightfn )
mc_weight( dst, i_dst_stride, src1, i_src_stride, weight, i_width, i_height );
else
mc_copy( src1, i_src_stride, dst, i_dst_stride, i_width, i_height );
}
static uint8_t *get_ref( uint8_t *dst, int *i_dst_stride,
uint8_t *src[4], int i_src_stride,
int mvx, int mvy,
int i_width, int i_height, const x264_weight_t *weight )
{
int qpel_idx = ((mvy&3)<<2) + (mvx&3);
int offset = (mvy>>2)*i_src_stride + (mvx>>2);
uint8_t *src1 = src[hpel_ref0[qpel_idx]] + offset + ((mvy&3) == 3) * i_src_stride;
if( qpel_idx & 5 ) /* qpel interpolation needed */
{
uint8_t *src2 = src[hpel_ref1[qpel_idx]] + offset + ((mvx&3) == 3);
pixel_avg( dst, *i_dst_stride, src1, i_src_stride,
src2, i_src_stride, i_width, i_height );
if( weight->weightfn )
mc_weight( dst, *i_dst_stride, dst, *i_dst_stride, weight, i_width, i_height );
return dst;
}
else if( weight->weightfn )
{
mc_weight( dst, *i_dst_stride, src1, i_src_stride, weight, i_width, i_height );
return dst;
}
else
{
*i_dst_stride = i_src_stride;
return src1;
}
}
/* full chroma mc (ie until 1/8 pixel)*/
static void mc_chroma( uint8_t *dst, int i_dst_stride,
uint8_t *src, int i_src_stride,
int mvx, int mvy,
int i_width, int i_height )
{
uint8_t *srcp;
int d8x = mvx&0x07;
int d8y = mvy&0x07;
int cA = (8-d8x)*(8-d8y);
int cB = d8x *(8-d8y);
int cC = (8-d8x)*d8y;
int cD = d8x *d8y;
src += (mvy >> 3) * i_src_stride + (mvx >> 3);
srcp = &src[i_src_stride];
for( int y = 0; y < i_height; y++ )
{
for( int x = 0; x < i_width; x++ )
dst[x] = ( cA*src[x] + cB*src[x+1] + cC*srcp[x] + cD*srcp[x+1] + 32 ) >> 6;
dst += i_dst_stride;
src = srcp;
srcp += i_src_stride;
}
}
#define MC_COPY(W) \
static void mc_copy_w##W( uint8_t *dst, int i_dst, uint8_t *src, int i_src, int i_height ) \
{ \
mc_copy( src, i_src, dst, i_dst, W, i_height ); \
}
MC_COPY( 16 )
MC_COPY( 8 )
MC_COPY( 4 )
void x264_plane_copy_c( uint8_t *dst, int i_dst,
uint8_t *src, int i_src, int w, int h)
{
while( h-- )
{
memcpy( dst, src, w );
dst += i_dst;
src += i_src;
}
}
static void prefetch_fenc_null( uint8_t *pix_y, int stride_y,
uint8_t *pix_uv, int stride_uv, int mb_x )
{}
static void prefetch_ref_null( uint8_t *pix, int stride, int parity )
{}
static void memzero_aligned( void * dst, int n )
{
memset( dst, 0, n );
}
static void integral_init4h( uint16_t *sum, uint8_t *pix, int stride )
{
int v = pix[0]+pix[1]+pix[2]+pix[3];
for( int x = 0; x < stride-4; x++ )
{
sum[x] = v + sum[x-stride];
v += pix[x+4] - pix[x];
}
}
static void integral_init8h( uint16_t *sum, uint8_t *pix, int stride )
{
int v = pix[0]+pix[1]+pix[2]+pix[3]+pix[4]+pix[5]+pix[6]+pix[7];
for( int x = 0; x < stride-8; x++ )
{
sum[x] = v + sum[x-stride];
v += pix[x+8] - pix[x];
}
}
static void integral_init4v( uint16_t *sum8, uint16_t *sum4, int stride )
{
for( int x = 0; x < stride-8; x++ )
sum4[x] = sum8[x+4*stride] - sum8[x];
for( int x = 0; x < stride-8; x++ )
sum8[x] = sum8[x+8*stride] + sum8[x+8*stride+4] - sum8[x] - sum8[x+4];
}
static void integral_init8v( uint16_t *sum8, int stride )
{
for( int x = 0; x < stride-8; x++ )
sum8[x] = sum8[x+8*stride] - sum8[x];
}
void x264_frame_init_lowres( x264_t *h, x264_frame_t *frame )
{
uint8_t *src = frame->plane[0];
int i_stride = frame->i_stride[0];
int i_height = frame->i_lines[0];
int i_width = frame->i_width[0];
// duplicate last row and column so that their interpolation doesn't have to be special-cased
for( int y = 0; y < i_height; y++ )
src[i_width+y*i_stride] = src[i_width-1+y*i_stride];
memcpy( src+i_stride*i_height, src+i_stride*(i_height-1), i_width+1 );
h->mc.frame_init_lowres_core( src, frame->lowres[0], frame->lowres[1], frame->lowres[2], frame->lowres[3],
i_stride, frame->i_stride_lowres, frame->i_width_lowres, frame->i_lines_lowres );
x264_frame_expand_border_lowres( frame );
memset( frame->i_cost_est, -1, sizeof(frame->i_cost_est) );
for( int y = 0; y < h->param.i_bframe + 2; y++ )
for( int x = 0; x < h->param.i_bframe + 2; x++ )
frame->i_row_satds[y][x][0] = -1;
for( int y = 0; y <= !!h->param.i_bframe; y++ )
for( int x = 0; x <= h->param.i_bframe; x++ )
frame->lowres_mvs[y][x][0][0] = 0x7FFF;
}
static void frame_init_lowres_core( uint8_t *src0, uint8_t *dst0, uint8_t *dsth, uint8_t *dstv, uint8_t *dstc,
int src_stride, int dst_stride, int width, int height )
{
for( int y = 0; y < height; y++ )
{
uint8_t *src1 = src0+src_stride;
uint8_t *src2 = src1+src_stride;
for( int x = 0; x<width; x++ )
{
// slower than naive bilinear, but matches asm
#define FILTER(a,b,c,d) ((((a+b+1)>>1)+((c+d+1)>>1)+1)>>1)
dst0[x] = FILTER(src0[2*x ], src1[2*x ], src0[2*x+1], src1[2*x+1]);
dsth[x] = FILTER(src0[2*x+1], src1[2*x+1], src0[2*x+2], src1[2*x+2]);
dstv[x] = FILTER(src1[2*x ], src2[2*x ], src1[2*x+1], src2[2*x+1]);
dstc[x] = FILTER(src1[2*x+1], src2[2*x+1], src1[2*x+2], src2[2*x+2]);
#undef FILTER
}
src0 += src_stride*2;
dst0 += dst_stride;
dsth += dst_stride;
dstv += dst_stride;
dstc += dst_stride;
}
}
#if defined(__GNUC__) && (defined(ARCH_X86) || defined(ARCH_X86_64)) && !defined(SPEC)
// gcc isn't smart enough to use the "idiv" instruction
static ALWAYS_INLINE int32_t div_64_32(int64_t x, int32_t y)
{
int32_t quotient, remainder;
asm("idiv %4"
:"=a"(quotient), "=d"(remainder)
:"a"((uint32_t)x), "d"((int32_t)(x>>32)), "r"(y)
);
return quotient;
}
#else
#define div_64_32(x,y) ((x)/(y))
#endif
/* Estimate the total amount of influence on future quality that could be had if we
* were to improve the reference samples used to inter predict any given macroblock. */
static void mbtree_propagate_cost( int *dst, uint16_t *propagate_in, uint16_t *intra_costs,
uint16_t *inter_costs, uint16_t *inv_qscales, int len )
{
for( int i = 0; i < len; i++ )
{
int propagate_amount = propagate_in[i] + ((intra_costs[i] * inv_qscales[i] + 128)>>8);
dst[i] = div_64_32((int64_t)propagate_amount * (intra_costs[i] - (inter_costs[i] & LOWRES_COST_MASK)), intra_costs[i]);
}
}
void x264_mc_init( int cpu, x264_mc_functions_t *pf )
{
pf->mc_luma = mc_luma;
pf->get_ref = get_ref;
pf->mc_chroma = mc_chroma;
pf->avg[PIXEL_16x16]= pixel_avg_16x16;
pf->avg[PIXEL_16x8] = pixel_avg_16x8;
pf->avg[PIXEL_8x16] = pixel_avg_8x16;
pf->avg[PIXEL_8x8] = pixel_avg_8x8;
pf->avg[PIXEL_8x4] = pixel_avg_8x4;
pf->avg[PIXEL_4x8] = pixel_avg_4x8;
pf->avg[PIXEL_4x4] = pixel_avg_4x4;
pf->avg[PIXEL_4x2] = pixel_avg_4x2;
pf->avg[PIXEL_2x4] = pixel_avg_2x4;
pf->avg[PIXEL_2x2] = pixel_avg_2x2;
pf->weight = x264_mc_weight_wtab;
pf->offsetadd = x264_mc_weight_wtab;
pf->offsetsub = x264_mc_weight_wtab;
pf->weight_cache = x264_weight_cache;
pf->copy_16x16_unaligned = mc_copy_w16;
pf->copy[PIXEL_16x16] = mc_copy_w16;
pf->copy[PIXEL_8x8] = mc_copy_w8;
pf->copy[PIXEL_4x4] = mc_copy_w4;
pf->plane_copy = x264_plane_copy_c;
pf->hpel_filter = hpel_filter;
pf->prefetch_fenc = prefetch_fenc_null;
pf->prefetch_ref = prefetch_ref_null;
pf->memcpy_aligned = memcpy;
pf->memzero_aligned = memzero_aligned;
pf->frame_init_lowres_core = frame_init_lowres_core;
pf->integral_init4h = integral_init4h;
pf->integral_init8h = integral_init8h;
pf->integral_init4v = integral_init4v;
pf->integral_init8v = integral_init8v;
pf->mbtree_propagate_cost = mbtree_propagate_cost;
#ifdef HAVE_MMX
x264_mc_init_mmx( cpu, pf );
#endif
#ifdef HAVE_ALTIVEC
if( cpu&X264_CPU_ALTIVEC )
x264_mc_altivec_init( pf );
#endif
#ifdef HAVE_ARMV6
x264_mc_init_arm( cpu, pf );
#endif
}
void x264_frame_filter( x264_t *h, x264_frame_t *frame, int mb_y, int b_end )
{
const int b_interlaced = h->sh.b_mbaff;
const int stride = frame->i_stride[0] << b_interlaced;
const int width = frame->i_width[0];
int start = (mb_y*16 >> b_interlaced) - 8; // buffer = 4 for deblock + 3 for 6tap, rounded to 8
int height = ((b_end ? frame->i_lines[0] : mb_y*16) >> b_interlaced) + 8;
int offs = start*stride - 8; // buffer = 3 for 6tap, aligned to 8 for simd
if( mb_y & b_interlaced )
return;
for( int y = 0; y <= b_interlaced; y++, offs += frame->i_stride[0] )
{
h->mc.hpel_filter(
frame->filtered[1] + offs,
frame->filtered[2] + offs,
frame->filtered[3] + offs,
frame->plane[0] + offs,
stride, width + 16, height - start,
h->scratch_buffer );
}
/* generate integral image:
* frame->integral contains 2 planes. in the upper plane, each element is
* the sum of an 8x8 pixel region with top-left corner on that point.
* in the lower plane, 4x4 sums (needed only with --partitions p4x4). */
if( frame->integral )
{
if( start < 0 )
{
memset( frame->integral - PADV * stride - PADH, 0, stride * sizeof(uint16_t) );
start = -PADV;
}
if( b_end )
height += PADV-9;
for( int y = start; y < height; y++ )
{
uint8_t *pix = frame->plane[0] + y * stride - PADH;
uint16_t *sum8 = frame->integral + (y+1) * stride - PADH;
uint16_t *sum4;
if( h->frames.b_have_sub8x8_esa )
{
h->mc.integral_init4h( sum8, pix, stride );
sum8 -= 8*stride;
sum4 = sum8 + stride * (frame->i_lines[0] + PADV*2);
if( y >= 8-PADV )
h->mc.integral_init4v( sum8, sum4, stride );
}
else
{
h->mc.integral_init8h( sum8, pix, stride );
if( y >= 8-PADV )
h->mc.integral_init8v( sum8-8*stride, stride );
}
}
}
}

60
src/speccpu/x264.py
View File

@@ -1,3 +1,4 @@
import importlib.resources
import os
import shutil
from pathlib import Path
@@ -8,6 +9,49 @@ def get_mc_path(build):
return build / "x264_src" / "common" / "mc.c"
class MCBuilder:
@staticmethod
def get_src():
assert __package__ is not None
with importlib.resources.open_text(__package__, "mc.c") as f:
mc_src = f.read()
return mc_src
def __init__(self):
self.mc_src = self.get_src()
def with_print_weight(self):
"""
Insert a printf statement after the opening brace in the mc_weight function.
"""
target_declaration = "static inline void mc_weight"
printf_statement = ' printf("weight->i_scale = %d\\n", weight->i_scale);'
# Find the position of the target function declaration
start_index = self.mc_src.find(target_declaration)
if start_index == -1:
raise ValueError("Could not find target function `mc_weight`")
# Find the position of the opening brace { after the target function declaration
open_brace_index = self.mc_src.find("{", start_index)
if open_brace_index == -1:
raise ValueError("Could not find opening brace { for the target function")
# Insert printf statement after the opening brace
self.mc_src = (
self.mc_src[: open_brace_index + 1] # Original content up to {
+ "\n"
+ printf_statement # Insert printf statement
+ "\n"
+ self.mc_src[open_brace_index + 1 :] # Remaining content
)
return self
def build(self):
return self.mc_src
def get_ref_add_pragma(mc_lines):
# Initialize variables
modified_lines = []
@@ -83,11 +127,8 @@ def recompile_mc(mc_path: Path, build: Path, spec_env: os._Environ):
# Run "make"
return make(build, spec_env)
def perf_ref(x264_run, perf_output):
"""Run "perf" to profile x264 reference performance."""
x264_cmd = [
x264_run / "x264_s_base.mytest-m64",
specinvoke_args = {
"ref": [
"--seek",
"500",
"--dumpyuv",
@@ -99,6 +140,15 @@ def perf_ref(x264_run, perf_output):
"BuckBunny.yuv",
"1280x720",
]
}
def perf_ref(x264_run, perf_output):
"""Run "perf" to profile x264 reference performance."""
x264_cmd = [
x264_run / "x264_s_base.mytest-m64",
*specinvoke_args["ref"],
]
run(
[
"perf",