From: user5857@newsgrouper.org.invalid   
      
   Michael S  posted:   
      
   > On Fri, 19 Dec 2025 03:30:26 -0000 (UTC)   
   > John Levine  wrote:   
   >   
   > > According to John Savard  :   
   > > >Therefore, I came up with the idea of using some opcode space for   
   > > >block headers which could contain information about the lengths of   
   > > >instructions, so as to make decoding variable-length instructions   
   > > >fully non-serialized, thus giving me the best of both worlds.   
   > >   
   > > Sounds like the first two bits of the opcode in S/360 which tells you   
   > > the instruction format which also tells you how long the instruction   
   > > is.   
   > >   
   > > They've added lots of new instructions since then with somewhat   
   > > different formats, but those bits still tell you how long the   
   > > instruction is. The first byte tells you what the fornat is so you   
   > > know what address calculations to do.   
   > >   
   > >   
   >   
   > With very long pipelines that IBM is using starting from z10 (17 years   
   > ago) it probably makes no difference.   
   > The fact that there are only 3 options for instruction length is   
   > important and simplifying things relatively to more than dozen of   
   > options in x86, but how many bits one has to access in order to   
   > determine the length of instruction is irrelevant or close to   
   > irrelevant as long as they all reside near beginning rather than   
   > anywhere, like in VAX.   
      
   My VLI stuff is all encoded in the first word (32-bits) of the instruction.   
   And (now) all extensions come in 32-bit quanta, and its all encoded in   
   4-bits.   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)