comparison libpostproc/postprocess_template.c @ 2041:b996fbe0a7e7 libavcodec

Newer version, using a vectorized version of the new organisation of code in doVertLowPass. it seems to be faster in AltiVec also... Also includes a compile fix for the new do_a_deblock when using AltiVec. patch by (Romain Dolbeau <dolbeau at irisa dot fr>)

2040:5de466b3360e

	);
}
#endif
//static int test=0;

static inline void RENAME(tempNoiseReducer)(uint8_t *src, int stride,
				    uint8_t *tempBlured, uint32_t *tempBluredPast, int *maxNoise)
{
	// to save a register (FIXME do this outside of the loops)
	tempBluredPast[127]= maxNoise[0];

//printf("%d\n", test);
#else
{
	int y;
	int d=0;
	int sysd=0;
	int i;

	for(y=0; y<8; y++)
	{
		int x;

			int d1=ref - cur;
//			if(x==0 || x==7) d1+= d1>>1;
//			if(y==0 || y==7) d1+= d1>>1;
//			d+= ABS(d1);
			d+= d1*d1;
			sysd+= d1;
		}
	}
	i=d;
	d= 	(
		4*d

		}
	}
}
#endif
}

#ifdef HAVE_MMX
/**
 * accurate deblock filter
 */

2041:b996fbe0a7e7

	);
}
#endif
//static int test=0;

#ifndef HAVE_ALTIVEC
static inline void RENAME(tempNoiseReducer)(uint8_t *src, int stride,
				    uint8_t *tempBlured, uint32_t *tempBluredPast, int *maxNoise)
{
	// to save a register (FIXME do this outside of the loops)
	tempBluredPast[127]= maxNoise[0];

//printf("%d\n", test);
#else
{
	int y;
	int d=0;
//	int sysd=0;
	int i;

	for(y=0; y<8; y++)
	{
		int x;

			int d1=ref - cur;
//			if(x==0 || x==7) d1+= d1>>1;
//			if(y==0 || y==7) d1+= d1>>1;
//			d+= ABS(d1);
			d+= d1*d1;
//			sysd+= d1;
		}
	}
	i=d;
	d= 	(
		4*d

		}
	}
}
#endif
}
#endif //HAVE_ALTIVEC

#ifdef HAVE_MMX
/**
 * accurate deblock filter
 */