home bbs files messages ]

Forums before death by AOL, social media and spammers... "We can't have nice things"

   comp.arch      Apparently more than just beeps & boops      131,241 messages   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]

   Message 130,508 of 131,241   
   Anton Ertl to MitchAlsup   
   Re: Memory ordering (Re: Multi-precision   
   13 Dec 25 09:31:05   
   
   From: anton@mips.complang.tuwien.ac.at   
      
   MitchAlsup  writes:   
   >What my solution entails is a modification   
   >to the cache coherence model (NaK) that indicates "Yes I have the line you   
   >referenced, but, no you can't have it right now" in order to strengthen   
   >the guarantees of forward progress.   
      
   How does it strengthen the guarantees of forward progress?  My guess:   
   If the requester itself is in an atomic sequence B, it will cancel it.   
   This could help if the atomic sequence A that caused the NaK then   
   tries to get a cache line that would be kept by B.   
      
   There is still a chance of both sequences canceling each other by   
   sending NaKs at the same time, but it is smaller and with something   
   like exponential backoff eventual forward progress could be achieved.   
      
   - 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)   

[   << oldest   |   < older   |   list   |   newer >   |   newest >>   ]


(c) 1994,  bbs@darkrealms.ca