NetNews Usenet Archive 1992 #16/NN_1992_16.iso/spool/comp/arch/8350

Text File | 1992-07-27 | 1.6 KB | 34 lines

Newsgroups: comp.arch
Path: sparky!uunet!elroy.jpl.nasa.gov!decwrl!adobe!usenet
From: [e-mail] (Zalman Stern)
Subject: Re: 64-bit CPU vs 2 x 32-bit CPUs
Message-ID: <[e-mail]>
Sender: [e-mail] (USENET NEWS)
Organization: Adobe Systems Incorporated
References: <[e-mail]>
Date: Mon, 27 Jul 1992 16:27:34 GMT
Lines: 22

In article <[e-mail]> [e-mail] (Ed Gould)
writes:
> > Since the DEC 21064 Alpha chip only supports 43-bit effective virtual
> > addresses, the cost is approximately 43-32=11 pins.
>
> According to the data sheet I have, the 21064 implements 43-bit virtual
> addresses (even though all 64 bits are checked), but only 34-bit
> physical addresses.  Hence, it's only *two* pins.  I can't imagine that
> that's particularly significant.
..And if I recall, the R4000 implements 40 bits of virtual and 36 bits of
physical. Go figure... One reason these chips do this is because at straight
32 bit virtual and physical their "high performance" nature becomes suspect.
That is, its a product differentiation feature. And for some things its, a
very important. For others its useless. I saw this disscussion during the
16/32 bit microprocessor transition and I'm sure it came up during the 8/16
bit one as well. A big difference now may be commercial multiprocessor
software existing for microprocessor architectures. (And one of the most
important pieces of such software probably won't go to 64 bits for a good
while.)
--
Zalman Stern           [e-mail]            [phone]
Adobe Systems, 1585 Charleston Rd., POB 7900, Mountain View, CA 94039-7900