Path: utzoo!utgpu!watmath!att!dptg!rutgers!usc!ucsd!ucbvax!tut.cis.ohio-state.edu!gem.mps.ohio-state.edu!ginosko!husc6!paperboy!pmax7!lehotsky
From: lehotsky@pmax7.osf.org (Alan Lehotsky)
Newsgroups: comp.arch
Subject: Re: Memory utilization & inter-process contention
Message-ID: <682@paperboy.OSF.ORG>
Date: 30 Aug 89 19:27:59 GMT
References: <3332@blake.acs.washington.edu> <261500008@S34.Prime.COM> <249@gp.govt.nz> <2389@auspex.auspex.com> <AGLEW.89Aug29104841@chant.urbana.mcd.mot.com> <2396@auspex.auspex.com>
Sender: news@OSF.ORG
Reply-To: lehotsky@pmax7.osf.org (Alan Lehotsky)
Organization: Open Software Foundation
Lines: 12


There is one other trick used by VMS (at least on the 780).  The original
11/780 did not have a translation-buffer "read-out".  So, it was
eco-ed to give access to the last 32 virtual-address translations.  This
gave them the most-recently used pages so that they could avoid a truely
random page-replacement strategy.

I don't know whether newer processors retained this feature - but I think
that it was an extremely clever trick!

Al Lehotsky
(former) Bliss Wizard