From: johnl@taugh.com   
      
   According to Anton Ertl :   
   >Stefan Monnier  writes:   
   >>> At the time of conception, there were amny arguments that {sooner or   
   >>> later} compilers COULD figure stuff like this out.   
   >>   
   >>I can't remember seeing such arguments comping from compiler people, tho.   
   >   
   >Actually, the IA-64 people could point to the work on VLIW (in   
   >particular, Multiflow (trace scheduling) and Cydrome (software   
   >pipelining)), which in turn is based on the work on compilers for   
   >microcode.   
      
   I knew the Multiflow people pretty well when I was at Yale. Trace   
   scheduling was inspired by the FPS AP-120B, which had wide   
   instructions issuing multiple operations and was extremely hard to   
   program efficiently.   
      
   Multiflow's compiler worked pretty well and did a good job of static   
   scheduling memory operations when the access patterns weren't too data   
   dependent. It was good enough that Intel and HP both licensed it and   
   used it in their VLIW projects.  It was a good match for the hardware   
   in the 1980s.   
      
   But as computer hardware got faster and denser, it became possible to   
   do the scheduling on the fly in hardware, so you could get comparable   
   performance with conventional instruction sets in a microprocessor.   
      
      
   --   
   Regards,   
   John Levine, johnl@taugh.com, Primary Perpetrator of "The Internet for   
   Dummies",   
   Please consider the environment before reading this e-mail. https://jl.ly   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)