Path: utzoo!telly!attcan!ncrcan!scocan!seanf
From: seanf@sco.COM (Sean Fagan)
Newsgroups: comp.arch
Subject: Re: 64 bits for times....
Message-ID: <1990Aug23.022416.14798@sco.COM>
Date: 23 Aug 90 06:24:16 GMT
References: <5539@darkstar.ucsc.edu> <13285@yunexus.YorkU.CA> <30728@super.ORG> <26012@bellcore.bellcore.com> <11187@alice.UUCP>
Reply-To: seanf@sco.COM (Sean Fagan)
Organization: The Santa Cruz Operation, Inc.
Lines: 25

In article <26012@bellcore.bellcore.com>, mo@messy.bellcore.com (Michael O'Dell) writes:
> One of the nice things about using 64 bits for the time is that
> you can then put it in nanoseconds - which you *almost* really
> need on really fast machines. (100ns might be ok, but the
> difference is still contained in 64 bits, so just do it!!)

The Elxsi, a rather nice machine (it has no supervisor mode!), has a 50ns
clock, and a 64-bit clock register.  If you want to find out *exactly* how
many clock-ticks an instruction takes, you do something like:

	ld.l	r1, CLOCK
	<instr>
	ld.l	r2, CLOCK
	sub.l	r3, r1, r2

I'm guessing at the syntax, and I'm *sure* it's wrong, but you get the
general idea.

Rather useful, actually.

-- 
Sean Eric Fagan  | "let's face it, finding yourself dead is one 
seanf@sco.COM    |   of life's more difficult moments."
uunet!sco!seanf  |   -- Mark Leeper, reviewing _Ghost_
(408) 458-1422   | Any opinions expressed are my own, not my employers'.