From: krishna.myneni@ccreweb.org   
      
   On 4/14/24 21:07, dxf wrote:   
   ...   
   > The design criteria that never changed was the 8-level hardware stack.   
   > Forthers can either accept it for best performance - or pick something   
   > more forgiving at a lesser performance.   
   >   
      
   In the Lorenz equation example, which works with the 8 deep fpu stack,   
   we have assumed that the fpu hardware stack was empty before calling   
   DERIVS. In a real use case, the call to DERIVS is likely to occur within   
   a deeper call chain, resulting in items already on the fpu stack before   
   args for DERIVS are pushed. As Marcel said, using only a hardware-based   
   fp stack is not realistic for any non-trivial floating point work.   
      
   The loss of performance with a memory-based fp stack is far less a   
   concern than having to consider the limited stack depth when writing   
   code involving floating point arithmetic. Failure from overflowing the   
   fpu stack is silent. Debugging is likely to be a nightmare.   
      
   --   
   Krishna   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)