Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!cbosgd!ihnp4!drutx!mtuxo!houxm!whuxl!whuxlm!akgua!gatech!seismo!rlgvax!hadron!jsdy
From: jsdy@hadron.UUCP
Newsgroups: net.micro.att
Subject: Re: 1meg vs 2meg 7300
Message-ID: <236@hadron.UUCP>
Date: Tue, 4-Feb-86 23:52:07 EST
Article-I.D.: hadron.236
Posted: Tue Feb  4 23:52:07 1986
Date-Received: Fri, 7-Feb-86 21:34:28 EST
References: <252@spp3.UUCP> <222@imagen.UUCP> <217@hoqam.UUCP>
Reply-To: jsdy@hadron.UUCP (Joseph S. D. Yao)
Distribution: net
Organization: Hadron, Inc., Fairfax, VA
Lines: 20
Summary: "swapping" vs "paging"; "thrashing" under either.

In article <217@hoqam.UUCP> vedm@hoqam.UUCP (BEATTIE) writes:
> [ attribution MISSING! ]
>> You are experiencing thrashing of memory pages - the system does not use
>> swapping, it has virtual memory.
>Please explain the distinction between "thrashing" and "swapping".

Not a "distinction" but an attribute.  Swapping is putting a whole
program out at a time from primary storage to secondary storage.
Paging, on the other hand, is putting a part out at a time:  the
advantage is that you can have the active part of more programs in
main memory at a time (theoretically).  When the system becomes so
swap-bound or page-bound that it is spending an inordinate amount
of time doing I/O for these functions, it is called "thrashing."
("Inordinate" can be 20-60%, depending on other factors.)

Most paging machines still swap some of the time.  I haven't looked
at s5r2.0 hard enough yet to determine whether this is true of it.
-- 

	Joe Yao		hadron!jsdy@seismo.{CSS.GOV,ARPA,UUCP}