[continued from previous message]   
      
   > If you want to be able to perform one taken branch per cycle (or   
   > more), you always need prediction.   
      
   Greater than 1 branch per FETCH latency.   
      
   > >If you use a link register or a special instruction, the CPU could   
   > >do that.   
   >   
   > It turns out that this does not work well in practice.   
      
   Agreed.   
      
   > - anton   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)