From: bc@nospicedham.freeuk.com   
      
   On 25/04/2018 03:52, Robert Wessel wrote:   
      
   > It's not code size, the operand for an and with 0xfffffff0 gets sign   
   > extended from one byte, so that's actually a shorted sequence.   
   > I wonder if somewhere in GCC's architecture definition for the   
   > platform the required rounding is defined as "((n+15)/16)*16".  The   
   > code would be a literal translation of that.   
   >   
   > OTOH, you really can't infer too much from GCC at -O0.  It generates   
   > really bad code.   
      
   No, really bad code would have used DIV and MUL (I'm an expert on that).   
      
   --   
   bartc   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)