Path: utzoo!attcan!uunet!zephyr.ens.tek.com!tekcrl!tekchips!kend
From: kend@tekchips.LABS.TEK.COM (Ken Dickey)
Newsgroups: comp.lang.lisp
Subject: Re: LISP compiler? (really future of MACL)
Summary: Real-time collector, cheap!
Message-ID: <5117@tekcrl.LABS.TEK.COM>
Date: 25 Nov 89 02:37:17 GMT
References: <5130@internal.Apple.COM> <5149@internal.Apple.COM> <616@pyuxf.UUCP> <2159@cs-spool.calgary.UUCP>
Sender: ftp@tekcrl.LABS.TEK.COM
Reply-To: kend@tekchips.LABS.TEK.COM (Ken Dickey)
Organization: Tektronix, Inc., Beaverton,  OR.
Lines: 18


> A real time (parallel or incremental) garbage collector is roughly
> twice as costly as a comparable stop-and-collect collector.  The real
> time version has to collect more frequently than stop-and-copy because
> the real time version has to start it's collection well before memory
> is exhausted.  A real time "two space copying" collector is even more
> expensive because of the cost of checking for a forwarding pointer on
> every reference to GC managed memory.  Real time collectors penalize
> real performance while improving some aspects of perceived
> performance.

This is not the case if you use your VM hardware.  See "Real-time
Concurrent Collection on Stock Multiprocessors" by Appel, Ellis, &
Lee; Princeton U. tech. report: CS-TR-133-88.  The method used also
allows multiple allocators.


-Ken Dickey		kend@mrloog.WR.TEK.COM