Use 3Dnow! Pavgusb For Mpeg-2 Motion Compensation - AMD Athlon Processor x86 Optimization Manual

22007E/0—November 1999
;;
;;
PXOR MM0, MM0
MOVQ MM1, MM6
MOVQ MM4, MM5
PUNPCKHDQ MM1, MM1
MOVQ MM3, MM6
MOVQ MM2, MM5
PFSUBR MM3, MM0
PFSUBR MM2, MM0
PUNPCKLDQ MM3, MM6
PFCMPGT MM4, MM1
MOVQ MM0, QWORD PTR [ABOVE_RIGHT]
PFCMPGT MM3, MM1
PFCMPGT MM2, MM1
MOVQ MM1, QWORD PTR [BEHIND_BEFORE] ; BEHIND | BEFORE
PAND MM4, MM0
MOVQ MM0, QWORD PTR [BELOW_LEFT]
PAND MM3, MM1
PAND MM2, MM0
POR MM2, MM4
POR MM2, MM3
MOVQ MM1, MM2
PUNPCKHDQ MM2, MM2
POR MM2, MM1
Use 3DNow!™ PAVGUSB for MPEG-2 Motion Compensation
Use 3DNow!™ PAVGUSB for MPEG-2 Motion Compensation
DESTROYS MM0,MM1,MM2,MM3,MM4
Use the 3DNow! PAVGUSB instruction for MPEG-2 motion
compensation. The PAVGUSB instruction produces the rounded
averages of the eight unsigned 8-bit integer values in the source
operand (a MMX register or a 64-bit memory location) and the
eight corresponding unsigned 8-bit integer values in the
destination operand (a MMX register). The PAVGUSB
instruction is extremely useful in DVD (MPEG-2) decoding
where motion compensation performs a lot of byte averaging
between and within macroblocks. The PAVGUSB instruction
helps speed up these operations. In addition, PAVGUSB can
free up some registers and make unrolling the averaging loops
possible.
The following code fragment uses original MMX code to
perform averaging between the source macroblock and
destination macroblock:
AMD Athlon™ Processor x86 Code Optimization
; 0 | 0
; w | z
; y | x
; w | w
; w | z
; y | x
; -w | -z
; -y | -x
; z | -z
; y>w?FFFFFFFF:0 | x>w?FFFFFFFF:0
; ABOVE | RIGHT
; z>w?FFFFFFFF:0 | -z>w>FFFFFFFF:0
; -y>w?FFFFFFFF:0 | -x>w?FFFFFFFF:0
; y > w ? ABOVE:0
; BELOW | LEFT
; z > w ? BEHIND:0 | -z > w ? BEFORE:0
; -y > w ? BELOW:0 | -x > w ? LEFT:0
; BELOW,ABOVE | LEFT,RIGHT
; BELOW,ABOVE,BEHIND | LEFT,RIGHT,BEFORE
; BELOW,ABOVE,BEHIND | LEFT,RIGHT,BEFORE
; BELOW,ABOVE,BEHIND | BELOW,ABOVE,BEHIND
; zclip, yclip, xclip = clip code
| x > w ? RIGHT:0
123