Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!mailrus!csd4.csd.uwm.edu!uxc.cso.uiuc.edu!dino!atanasoff!hascall
From: hascall@atanasoff.cs.iastate.edu (John Hascall)
Newsgroups: comp.arch
Subject: Re: Working Set replacement, use bits (was re: memory utilization...)
Keywords: working set, replacement, virtual memory, use bits
Message-ID: <1402@atanasoff.cs.iastate.edu>
Date: 29 Aug 89 14:04:26 GMT
References: <11564@polya.Stanford.EDU>
Reply-To: hascall@atanasoff.cs.iastate.edu.UUCP (John Hascall)
Organization: Iowa State Univ. Computation Center
Lines: 23

In article <11564> wilson@carcoar.Stanford.EDU (Paul Wilson) writes:
 
}   An explanation of WS:
 
}The basic idea of the Working Set policy is that each process gets
}to keep in memory all of the pages that it has referenced within
}some period of time T.  For any given process, the Least Recently Used
}(LRU) page is always evicted.  What makes it different from normal
}LRU is that the amount of memory allocated to a process is variable --
}it is any number of pages referenced within the last T units of time.
}(T units of the given process' execution time that is.)
 
    It has been a while since I looked at this, so my memory could be
    a little moldy...

    I believe that when designing the WS scheme for VMS they studied
    various replacement algorithms (LRU, FIFO and RANDOM) and found
    that RANDOM was just about as effective as the other two.  It also
    has the virtue of being easier to implement.

    Does VMS, in fact, still use random replacement?

 John Hascall