optimization : arithmetic operations

[arakov]

Optimize the following bytecode snip:

       set dp   :-8
       store sp :0
       set dp   :-4
       store sp :1
       copy dp  :-16, :4
       xmov sp  :0, :1            
       iadd dp  :-16, :4
       set dp   :-16
       load     
       save dp  :-12

rearrange the operation, to get rid of xmov sp:0, :1