roy@phri.UUCP (Roy Smith) (04/16/85)
> [...] But computer > systems limit how much of it you can address. For example IBM's > MVS pre-XA uses 24 bit addressing so you can only use up to 16Mb > of virtual memory. Which by the way, is not enough for some of our > users. [...] I'm not sure why this is in net.puzzle, so I'm followup-to'ing it to net.arch, and cross-posting this there. I remember reading once that the biggest address size you would ever need would be 200 bits. The reason for this is that there are 2**200 atoms in the universe and it is absurd to want to enumerate more things than that. Any comments? BTW, Offhand I can already think of things like sparse hash tables which might fly in the face of this argument. -- cmcl2!rocky2!cubsvax -\ vax135!timeinc -> !phri!roy (Roy Smith, System Administrator) allegra -/ The opinions expressed herein do not necessarily reflect the views of the Public Health Research Institute.
jer@peora.UUCP (J. Eric Roskos) (04/16/85)
> I remember reading once that the biggest address size you > would ever need would be 200 bits. The reason for this is that there > are 2**200 atoms in the universe and it is absurd to want to enumerate > more things than that. Any comments? Certainly ... you are assuming that all computations involve things in the concrete, physical, discrete world; and that such computations require some enumeration of a subset of these objects; but this is clearly not the case. For instance, suppose you want to generate the set of 2**200+1 prime numbers, and keep them in a table so they'll be readily available. It's not even possible, I don't think, to argue that you'd run out of storage media, since your storage medium might be nonatomic; consider two pairs of antennae, positioned arbitrarily far apart, with one antenna sending a stream of bits to the remote antenna, which then reflects back the signal via the other antenna of the pair. Clearly this is a storage medium, and does not use any atoms to actually store the data; and if positioned arbitrarily far apart (assuming the signal would still be detectable at the remote site) could store arbitrarily much data. -- Full-Name: J. Eric Roskos UUCP: ..!{decvax,ucbvax,ihnp4}!vax135!petsd!peora!jer US Mail: MS 795; Perkin-Elmer SDC; 2486 Sand Lake Road, Orlando, FL 32809-7642
mhc@brunix.UUCP (Mark Copley) (04/19/85)
Summary: Expires: Sender: Followup-To: Distribution: Keywords: The use of two antennas located an arbitrary distance apart, one sending a bit stream and the other receiving it (a cosmic delay line), is an is an interesting idea, but wouldn't this storage media still be physically limited by the number of photons in the universe used to propagate the bit stream as an EM wave? Mark Copley
simon@psuvax1.UUCP (04/23/85)
Keywords:number of atoms in universe I can't resist. The assertion that arbitrary amounts of data (arbitrary as in more that the number of atoms in the universe or other similar quantities) could be stored by reflecting electromagnetic radiation between two antennas, far away from each other, is nonsense. In order to create and to mantain the field you use energy. There is only a fifinte amount of it (no, you cannot reduce field intensity: photon energies only decrease with wawelength, and you cannot have wawelengths that are too long relative to the device geometry).
mark@tove.UUCP (Mark Weiser) (04/23/85)
In article <202@phri.UUCP> roy@phri.UUCP (Roy Smith) writes: >... I remember reading once that the biggest address size you would >ever need would be 200 bits. The reason for this is that there are >2**200 atoms in the universe and it is absurd to want to enumerate more >things than that. Any comments? The number of atoms argument is silly. Prime numbers are things that people certainly want to enumerate, and their number is unbounded. -mark -- Spoken: Mark Weiser ARPA: mark@maryland Phone: +1-301-454-7817 CSNet: mark@umcp-cs UUCP: {seismo,allegra}!umcp-cs!mark USPS: Computer Science Dept., University of Maryland, College Park, MD 20742
phil@osiris.UUCP (Philip Kos) (04/24/85)
> . . . . consider two pairs > of antennae, positioned arbitrarily far apart, with one antenna sending a > stream of bits to the remote antenna, which then reflects back the signal > via the other antenna of the pair. Clearly this is a storage medium, and > does not use any atoms to actually store the data; and if positioned > arbitrarily far apart (assuming the signal would still be detectable at the > remote site) could store arbitrarily much data. This is an interesting idea, but sounds a bit tough to implement. A more realistic variation might involve sending the data over a line which imposes an arbitrarily long delay. To this end, how about using USENET, or, better yet, USNAIL (of course, in either of these cases, you wouldn't want to worry about disappearing data)? (I just LOVE silliness. Reaaaaalllly, I do.) Phil Kos The Johns Hopkins Hospital ...!umcp-cs!aplvax!osiris!phil Message undeliverable: unknown site "a galaxy far, far away"
ron@brl-tgr.ARPA (Ron Natalie <ron>) (04/26/85)
> In theory, no you can't run out of virtual memory. But computer > systems limit how much of it you can address. For example IBM's > MVS pre-XA uses 24 bit addressing so you can only use up to 16Mb > of virtual memory. Which by the way, is not enough for some of our > users. That's why we are going to MVS/XA and 31 bit addressing. Ahem, we seem to be having some difficulty with the term virtual here. Virtual memory is the use of more memory than you have physically. If I have an 8K system with a 16K paging area, I run out of virtual memory at 24K (assuming an optimal solution). No where in the definition of virtual memory does it say "infinite" amounts are available. -Ron Probably one of those people who thought that the infinite speed motor controls allowed you to make the motors go infinitely fast.
shp@crystal.UUCP (04/26/85)
> > Ahem, we seem to be having some difficulty with the term virtual > here. Virtual memory is the use of more memory than you have physically. > > -Ron > I agree. We ARE having difficulty. Still. The key is the word "virtual." What it means is that there does NOT NECESSARILY exist a one-to-one relationship between logical addresses and physical addresses. Some virtual memories may have a good deal more phyisical memory than the address space allows (consider the PDP-11/70). I realize this point may not be germaine to the discussion at hand; just wanted to keep this straight (it may be important to someone....). =shp