Date: Thu, 11 Dec 1997 10:12:12 -0800 (PST) From: Tom Bartol <bartol@salk.edu> To: Jonathan Lemon <jlemon@americantv.com> Cc: "John S. Dyson" <toor@dyson.iquest.net>, "J. Weatherbee - Senior Systems Architect" <jamil@acroal.com>, hackers@FreeBSD.ORG Subject: Re: This IS relevant, you'll realize why later. Message-ID: <Pine.BSF.3.96.971211093559.3358C-100000@dale.salk.edu> In-Reply-To: <19971211090510.50776@right.PCS>
next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, 11 Dec 1997, Jonathan Lemon wrote: > Hmm. One of my professors was claiming that with 128 bits, you would have > an address space of 2^128, which would be large enough that you could do > away with virtual memory altogether. > > Just give each piece of data it's own unique address, (ala Multics) and > then never relocate it for the lifetime of the system. She pointed out > that even the human genome database project is smaller than this. > > Somehow, I don't quite buy the argument. > -- > Jonathan > I'd buy it for $5.00 -- If the logic for each bit fit in a cube 0.14 picometer on a side you could fit 2^128 bits in 1 cubic meter. Never mind that 0.14 picometer is ~1/700th the diameter of a hydrogen atom and is only about 100 times bigger than an atomic nucleus! So if we could fit the logic for one bit in the space of a hydrogen atom, 2^128 bits would fill a cube ~700 meters on a side! Or if one bit fit in a cube 0.25 microns on a side then 2^128 bits would fill a cube ~1750 kilometers on a side (i.e. about 1/4th the volume of the Moon)!!!! Memories like this should be available in our life-time, cost about $5 each, and you'll need a whole bunch of them to run the latest version of MS Word available at the time. :-0 Tom
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?Pine.BSF.3.96.971211093559.3358C-100000>