XPost: alt.folklore.computers   
   From: anton@mips.complang.tuwien.ac.at   
      
   Thomas Koenig  writes:   
   >De Castro had had a big success with a simple load-store   
   >architecture, the Nova.  He did that to reduce CPU complexity   
   >and cost, to compete with DEC and its PDP-8.  (Byte addressing   
   >was horrible on the Nova, though).   
      
   The PDP-8, and its 16-bit followup, the Nova, may be load/store, but   
   it is not a register machine nor byte-addressed, while the PDP-11 is,   
   and the RISC-VAX would be, too.   
      
   >Now, assume that, as a time traveler wanting to kick off an early   
   >RISC revolution, you are not allowed to reveal that you are a time   
   >traveler (which would have larger effects than just a different   
   >computer architecture).  What do you do?   
   >   
   >a) You go to DEC   
   >   
   >b) You go to Data General   
   >   
   >c) You found your own company   
      
   Even if I am allowed to reveal that I am a time traveler, that may not   
   help; how would I prove it?   
      
   Yes, convincing people in the mid-1970s to bet the company on RISC is   
   a hard sell, that's I asked for "a magic wand that would convince the   
   DEC management and workforce that I know how to design their next   
   architecture, and how to compile for it" in   
   <2025Mar1.125817@mips.complang.tuwien.ac.at>.   
      
   Some arguments that might help:   
      
   Complexity in CISC and how it breeds complexity elsewhere; e.g., the   
   interaction of having more than one data memory access per   
   instruction, virtual memory, and precise exceptions.   
      
   How the CDC 6600 achieved performance (pipelining) and how non-complex   
   its instructions are.   
      
   I guess I would read through RISC-vs-CISC literature before entering   
   the time machine in order to have some additional arguments.   
      
      
   Concerning your three options, I think it will be a problem in any   
   case.  Data General's first bet was on FHP, a microcoded machine with   
   user-writeable microcode, so maybe even more in the wrong direction   
   than VAX; I can imagine a high-performance OoO VAX implementation, but   
   for an architecture with exposed microcode like FHP an OoO   
   implementation would probably be pretty challenging.  The backup   
   project that eventually came through was also a CISC.   
      
   Concerning founding ones own company, one would have to convince   
   venture capital, and then run the RISC of being bought by one of the   
   big players, who buries the architecture.  And even if you survive,   
   you then have to build up the whole thing: production, marketing,   
   sales, software support, ...   
      
   In any case, the original claim was about the VAX, so of course the   
   question at hand is what DEC could have done instead.   
      
   - anton   
   --   
   'Anyone trying for "industrial quality" ISA should avoid undefined behavior.'   
     Mitch Alsup,    
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)