Commit 2a84512a authored by Michael Niedermayer's avatar Michael Niedermayer

Optimize horizontal_compose53i() by inlining inv_lift() and optimizing the

resulting code slightly. ~30% faster C 5/3 wavelet at the decoder side.

Originally committed as revision 21130 to svn://svn.ffmpeg.org/ffmpeg/trunk
parent 18ca491b
......@@ -1015,45 +1015,23 @@ static void horizontal_compose53i(IDWTELEM *b, int width){
const int w2= (width+1)>>1;
int x;
#if 0
int A1,A2,A3,A4;
A2= temp[1 ];
A4= temp[0 ];
A1= temp[0+width2];
A1 -= (A2 + A4)>>1;
A4 += (A1 + 1)>>1;
b[0+width2] = A1;
b[0 ] = A4;
for(x=1; x+1<width2; x+=2){
A3= temp[x+width2];
A4= temp[x+1 ];
A3 -= (A2 + A4)>>1;
A2 += (A1 + A3 + 2)>>2;
b[x+width2] = A3;
b[x ] = A2;
A1= temp[x+1+width2];
A2= temp[x+2 ];
A1 -= (A2 + A4)>>1;
A4 += (A1 + A3 + 2)>>2;
b[x+1+width2] = A1;
b[x+1 ] = A4;
}
A3= temp[width-1];
A3 -= A2;
A2 += (A1 + A3 + 2)>>2;
b[width -1] = A3;
b[width2-1] = A2;
#else
inv_lift(temp , b , b+w2, 1, 1, 1, width, 1, 2, 2, 0, 1);
inv_lift(temp+w2, b+w2, temp, 1, 1, 1, width, -1, 0, 1, 1, 1);
#endif /* 0 */
for(x=0; x<width2; x++){
b[2*x ]= temp[x ];
b[2*x + 1]= temp[x+w2];
temp[2*x ]= b[x ];
temp[2*x + 1]= b[x+w2];
}
if(width&1)
b[2*x ]= temp[x ];
temp[2*x ]= b[x ];
b[0] = temp[0] - ((temp[1]+1)>>1);
for(x=2; x<width-1; x+=2){
b[x ] = temp[x ] - ((temp[x-1] + temp[x+1]+2)>>2);
b[x-1] = temp[x-1] + ((b [x-2] + b [x ]+1)>>1);
}
if(width&1){
b[x ] = temp[x ] - ((temp[x-1]+1)>>1);
b[x-1] = temp[x-1] + ((b [x-2] + b [x ]+1)>>1);
}else
b[x-1] = temp[x-1] + b[x-2];
}
static void vertical_compose53iH0(IDWTELEM *b0, IDWTELEM *b1, IDWTELEM *b2, int width){
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment