Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!mit-eddie!ll-xn!ames!aurora!labrea!jade!eris!mwm
From: mwm@eris.BERKELEY.EDU (Mike (My watch has windows) Meyer)
Newsgroups: comp.arch
Subject: Re: register windows
Message-ID: <5567@jade.BERKELEY.EDU>
Date: Thu, 22-Oct-87 03:40:23 EDT
Article-I.D.: jade.5567
Posted: Thu Oct 22 03:40:23 1987
Date-Received: Sat, 24-Oct-87 16:45:34 EDT
References: <201@PT.CS.CMU.EDU> <8801@utzoo.UUCP> <8758@shemp.UCLA.EDU>
Sender: usenet@jade.BERKELEY.EDU
Reply-To: mwm@eris.BERKELEY.EDU (Mike (My watch has windows) Meyer)
Organization: Missionaria Phonibalonica
Lines: 42
Keywords: register windows, interrupt latency

In article <8758@shemp.UCLA.EDU> frazier@CS.UCLA.EDU (Greg Frazier) writes:
<In article <8801@utzoo.UUCP> henry@utzoo.UUCP (Henry Spencer) writes:
<>                                 Register windows may end up being a net win
<>despite context-switch overhead.
<>-- 

Every time I've looked at an article on register windows, I think of
something I refer to as "the Johnson stack machine." This machine
seemed to have all the advantages of register windows, and avoided
most of the disadvantages.

The basic idea was to not bother with either a data cache or a
register file. Instead, you used all that chip space for a
register-speed cache that is mapped onto memory. The mapping is tracked
by a base/length pair. It's a write-through cache.

The trick is that this is mapped onto the top of the stack. Pushing
something onto the stack involved incrementing the length counter,
possibly wrapping around in the file (my mind goes hazy on how that
worked - it may have actually required shuffling elements in the
cache).

But context switches were *cheap*. The save half involved nothing more
than making sure all writes had actually finished. Loading a new
context required reloading (tagging words as invalid, maybe?) the
file.

I have good reason to believe that AT&T was working on hardware that
used this model. Does anyone know what became of that?

Better yet, can someone who actually understands hardware explain why
I've not seen anything like the above in the real world?

	Thanx,
	<mike


--
I'm gonna lasso you with my rubberband lazer,		Mike Meyer
Pull you closer to me, and look right to the moon.	mwm@berkeley.edu
Ride side by side when worlds collide,			ucbvax!mwm
And slip into the Martian tide.				mwm@ucbjade.BITNET