From: blockedofcourse@foo.invalid   
      
   On 2/12/2026 3:39 PM, Lasse Langwadt wrote:   
   > so it had a bunch of orthogonal registers that could be used for anything you   
   > like, including stack pointer(s) though I don't see   
   > why you wouldn't always use R6   
   >   
   > that is register oriented, stack oriented is things like Forth and Java   
      
   Stack-oriented doesn't require use of THE stack but, rather, *A*   
   stack.  So, all you need is the ability to address memory through   
   a pointer to implement one.   
      
   [In the Z*80 family, EX (SP),HL was probably one of the most   
   cherished opcodes for this type of design approach]   
      
   It is not uncommon to have different stacks for different uses.   
      
   An implicit advantage is typically greater locality of reference   
   making caching potentially more effective.  And, instruction   
   sets tend to be simpler (cleaner?).   
      
   [Imagine what goes on "under the hood" when executing C++ code   
   vs LISP]   
      
   CPUs with register files are more likely to romp around memory   
   indiscriminately as you can easily switch between many "pointers"   
   to many different objects.  They typically have much more processor   
   state making them annoying for use with real-time activities   
   (unless they have "tricks" to making preserving state easier/faster)   
      
   The B5000 (et al.) were probably the earliest commercially successful   
   "native" stack machines -- and, in general, much more interesting than   
   the more banal designs that have followed.   
      
   Thankfully, processors are fast enough that you can *emulate* many   
   of these features instead of being stuck with the more mundane   
   mechanisms present.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)