Forgot your password?
typodupeerror

SKA Telescope To Provide a Billion PCs Worth of Processing 186

Posted by ScuttleMonkey
from the scaling-up-scalability dept.
Sharky2009 writes "IBM is researching an exaflop machine with the processing power of about one billion PCs. The machine will be used to help process the Exabyte of data per day expected to flow off the Square Kilometre Array (SKA) telescope project. The company is also researching solid state storage technology called 'racetrack memory' which is much faster and denser than flash and may hold the secret to storing the data from the SKA. The story also says that the SKA is unlikely to use grid computing or a cloud-based approach to processing the telescope data due to challenge in transferring so much data (about one thousand million 1Gb memory sticks each day)."
This discussion has been archived. No new comments can be posted.

SKA Telescope To Provide a Billion PCs Worth of Processing

Comments Filter:
  • by Lord Ender (156273) on Friday September 18, 2009 @04:28PM (#29471061) Homepage

    (about one thousand million 1Gb memory sticks each day)

    Could we get that in LoC's? Also, could we stick to the standard "one million thousands" unit, please?

    • Re: (Score:3, Funny)

      by Hadlock (143607)

      http://www.jamesshuggins.com/h/tek1/how_big.htm [jamesshuggins.com]

      So roughly 20 million Library of Congresses (20mm LoC)

      • by rubycodez (864176) on Friday September 18, 2009 @04:52PM (#29471381)

        could we please stick to serious measures of information within the field of IT instead of silly printer paper units, how many station wagons full of 9 track tape is that?

        • Which kind of station wagon? One the size of a Chevrolet Nomad or Oldsmobile Vista Cruiser or a small one like the Honda Fit? (If you know what 9 track tape is, then you know it's a relevant question.)
          • Are the seats being left in or are they taken out? What about the spare tire?
            • Ok, Vista Cruiser. Rear seat stowed. Front passenger seat left in, just in case you see a hot woman hitch hiking (even though it's statistically unlikely). Rear tire occupies space in side compartment behind rear wheel well, on side opposite of fuel tank -- we'll leave that because Murphy is a bitch. If we left it out that would be the one time a truck dumps a box of nails in the highway right in front of us, with the hot woman standing by the road just over the horizon 5 miles ahead...

              With the added

        • Re: (Score:2, Informative)

          by Phoenixlol (1549649)
          It's be roughly 105,882,352,941,176,470,588 discs.

          They are about 10.5 in in diameter and .5 in thick but encased for storage probably 12*1 which would be about 12 tapes per cubic foot

          The Volvo V70 has about 72 cubic feet of free space

          About 122,549,019,607,843,137 Volvo V70 Station Wagons... or one making the trip 122,549,019,607,843,137 times *shrug*
        • by Hadlock (143607)

          You'd need 58.82 million 9-track tapes. What are the dimensions of a 9 track tape, and what type of station wagon? I'm assuming 1964 vintage? Assuming 100 tapes per car, that's 5.8 million station wagons, or 1000 tapes per car, that's 588,000 station wagons. Are we including roof rack space?

          • Re:thousand million? (Score:5, Informative)

            by idontgno (624372) on Friday September 18, 2009 @06:20PM (#29472209) Journal

            Maximum diameter of 10.5" [wikipedia.org]. Tape width was 1/2". The protective ring around the tape reel would add about another 1/4" to the diameter, and the thickness of the reel sides and retaining ring would add about another 1/4" to the thickness.

            We're not going for a rigorous space-filling solution; we'll stack the tapes in a square array (rather than, say, a hexagonal one). So the tapes effectively become 10.75"^2 x .75" rectangular prisms. That's 0.0501573351 cubic feet per tape.

            According to this scan [wingedmessenger.net] of the 1972 Mercury station wagon brochure, the 1972 Montego MX had 91.6 cu ft of cargo space. That's 1826 tapes. (Assuming dimensions of tapes and station wagon are compatible and don't leave some wasted modulus.) So 1,000 tapes is in the right order of magnitude, but a smidge low. Good guess. But your "number of cars" division was off by a factor of ten. Using the correct version of your numbers, 58,820,000 tapes transported 1000 tapes at a time is 58,820 loads. Using the 1,826 tapes per load number, it works out to 32,213 loads.

            I calculate that the capacity of a station wagon full of 9-track tapes works out to 310 GB. (170 MB per late-era IBM 3400-series tape reel at maximum length, 32K blocking, and 6250BPI. At least that's what Wikipedia says.)

            Hmm... I wonder if the various IP performance calculations would work out for a MTU of 310GB and a ping time of minutes to hours (depending on trip length and freeway speeds)....

            • by muridae (966931)

              IP performance, as stated, is going to stink. Unlike solid-state memory over messenger pigeon, the transit time of tape over station wagon is going to be on the same order as the time needed to construct the packets. I am assuming a smart IP state machine, so reconstructing the station wagons will not be needed.

              Using the same IBM 3400 series tape, 170MB each, with a transfer speed of 1,250,000 B/s each tape would take roughly 142.6 seconds to write. Not counting changing the tapes or moving them into the ve

    • Re:thousand million? (Score:5, Informative)

      by rm999 (775449) on Friday September 18, 2009 @04:56PM (#29471437)

      A thousand million is probably the most correct term for international understanding.

      There is no world standard term for one thousand million. In the US and most of the UK we call it a "billion", but in several countries a billion means a million million. In these countries, a thousand million is usually called "a thousand million" or a "milliard", but I've never seen "a million thousand".

      • by afidel (530433) on Friday September 18, 2009 @05:03PM (#29471499)
        The long rule is stupid, if you are going to use units as big as a million million just use scientific notation.
      • by Lord Ender (156273) on Friday September 18, 2009 @05:05PM (#29471513) Homepage

        Here's the way it works:

        In the US and the UK, the number is officially called "billion." In India, it's called "100 Crore." Australia officially has no idea how they do their numbers, and Canada doesn't even know what language it speaks. There are no other English-speaking countries of consequence.

        Therefore, "billion" is the most acceptable term for international English-language writing.

        • Re: (Score:3, Insightful)

          by ceoyoyo (59147)

          I guess you're joking, but in Canada one thousand million is definitely one billion, in english. It is not in french (it is a millard), but in that case you'd write the rest of the article in french too.

        • by rdnetto (955205)

          Australia officially has no idea how they do their numbers

          I'm not sure about it being official, but everyone here uses 1 billion = 10^9

      • Re: (Score:3, Insightful)

        by Hurricane78 (562437)

        Why not Just say 1E9 or even 1 * 10^9 for "a thousand million". If someone has a problem with understanding that, then what does he do on this site anyways? ^^
        (Ok, actually everybody had this at school, so I can expect this to be a normal term, used on national television. But noo, they *could* lose the total retards by not using it. We can't have that!! :/)

        • by timeOday (582209)

          Why not Just say 1E9 or even 1 * 10^9 for "a thousand million".

          While we're at it, I'd prefer 10^9 ("ten to the 9th power") over 1*10^9 ("one times ten to the 9th power"). The "one times..." isn't accomplishing anything.

      • by guruevi (827432)

        You know they found a solution for that. It's called scientific or exponential notation. And once you learn it, it's quite simple - you just add the specified number of zero's or move the comma the correct number of times.

        • You know they found a solution for that. It's called scientific or exponential notation. And once you learn it, it's quite simple - you just add the specified number of zero's or move the comma the correct number of times.

          Not all countries use commas in their numbers.

      • Re: (Score:3, Insightful)

        by TheRaven64 (641858)
        Short billion is internationally understood too, and is the correct term for a thousand million. I've not seen a 1Gb memory stick for a while though, did the author mean 1GB? It's only a factor of eight difference, but it's important.
      • Here is what to do with international standards/systems when in doubt. If British didn't come up with their own weirdo standard, it is generally universal.

        Back in day before DVB-* became standard, we had a map of TV broadcasting systems (analogue) on a big World map, you could see there is a single horrible variation (in compatibility sense) of PAL, PAL's British variant which has same video spec but some really needless variation of sound spec. So, if you don't know about it and assume British PAL is the P

  • So is that the processing power of one billion IBM PC 5150s?
    • So is that the processing power of one billion IBM PC 5150s?

      IBM hasn't made PCs since 2005, when it sold its personal computing division to Lenovo. So it would be 1e9 of whatever computer Lenovo is selling.

  • Ska? (Score:3, Funny)

    by mcgrew (92797) * on Friday September 18, 2009 @04:37PM (#29471171) Homepage Journal

    Let's hear it for Reel Big Fish and the Pietasters! Is there a Reggae telescope?

  • What? (Score:4, Funny)

    by Anonymous Coward on Friday September 18, 2009 @04:39PM (#29471203)

    SKA telescope? Madness!

  • by Fulcrum of Evil (560260) on Friday September 18, 2009 @04:40PM (#29471213)
    Seriously, how is a PC a unit of processing ability? And one thousand million GB sticks is an Exabyte (hence the name). Perhaps you can just say 10^18 bytes. This is slashdot, not msnbc.
    • by clem.dickey (102292) on Friday September 18, 2009 @04:50PM (#29471349)
      Original article also compares a Peta of floating ops per *second* to an Exa of byte "processing and storing" per *day.* Journalism profs should save that article for class discussion.
      • by Chris Burke (6130)

        Journalism profs should save that article for class discussion.

        Yeah and they'll point at this article and say "See this? This is how you do it. Remember that your target audience has no idea what these egg-heads are saying, so it helps if you don't either. You can't spell 'dumbed down' without 'dumb'!"

      • Original article also compares a Peta of floating ops per *second* to an Exa of byte "processing and storing" per *day.* Journalism profs should save that article for class discussion.

        If a system uses one petaflops to process one exabyte per day, that's the same as saying it takes roughly 86,400 floating ops to process each 1000 bytes of data. That sounds not unreasonable.

        • by walshy007 (906710)

          actually, you may want to look into SI prefixes, exa is 1000 times peta, so what is being said is it takes 1e15 operations to compute 1e18 bytes of data.

          That does sound just a bit unreasonable

    • by NoYob (1630681) on Friday September 18, 2009 @04:51PM (#29471363)

      Seriously, how is a PC a unit of processing ability? And one thousand million GB sticks is an Exabyte (hence the name). Perhaps you can just say 10^18 bytes. This is slashdot, not msnbc.

      Some of us went and got an MBA; upon which, it knocked tens of points off of our IQ.

      Now, 10 carrots 18? 18 what? Rabbits?

      It should read 10 carrots and 18 rabbits!

      And people say I'm stewped!

    • Seriously, how is a PC a unit of processing ability? And one thousand million GB sticks is an Exabyte (hence the name). Perhaps you can just say 10^18 bytes. This is slashdot, not msnbc.

      Or even a billion GB for those of us not in the Commonwealth. Of course, not everyone knows how big an exabyte is, but a billion 1 GB memory sticks is a pretty good visualization.

      • Only US English uses billion to mean "1,000,000,000". In most other languages that use the word billion it means "1,000,000,000,000".

        And there are a hell of a lot more outside the Commonwealth than there are in the US.

    • by T Murphy (1054674) on Friday September 18, 2009 @04:59PM (#29471469) Journal
      He's just giving you a good comparison so you can picture an equivalent system using everyday objects. Kind of like how instead of just saying "Bob is tall", you should use strong imagery like "Bob is as tall as a 6-foot-4-inch tall tree"
    • Perhaps you can just say 10^18 bytes. This is slashdot, not msnbc

      Right, so how many tenths of a gagillion is that?

    • But your missing the point.. It was a thousand million 1GB memory sticks.

      if you were a true geek, you would have first asked if that was a 1 GB memory stick, or a 1GiB of capacity, then you would of asked how they were formatting the drives, to see how much usable space their really would be. When your talking about a thousand million of them.. the difference between 1024 and 1000 used in counting, plus the loss of area for formatting and such really ads up!
      • if you were a true geek, you would have first asked if that was a 1 GB memory stick, or a 1GiB of capacity,

        If you used that GiB thing much more, I'll have to stab you in the neck.

        /real geek

      • by smoker2 (750216)
        Considering TFS quotes it as gigabits, I'd say this was moot.
      • Re: (Score:3, Insightful)

        by TheRaven64 (641858)
        Actually, it was a 1Gb memory stick, which is 128MB. If you are a geek, you deny the existence of any prefix containing 'ibi'.
    • by timeOday (582209)

      Seriously, how is a PC a unit of processing ability?

      Of course, there is no correct (or even really meaningful) way to measure "processing ability" in general. Doubtless this machine's cores would be more like DSPs or GPU cores rather than CPU cores anyways. It would be insane to use general purpose CPUs for a specialized task on this scale.

    • Oh, come on. It's not a BS unit. It's not like we're trying to scientifically measure them, just convey some idea what these big numbers might mean in a more "real world" sense. And, in that sense, the cliche of "LOC's" actually carries some legitimacy, in that books contain data that can be easily compared to bytes. When somebody asks me what a "byte" is, I usually start by explaining that 1 byte is a keystroke. While not strictly true (Unicode can take up to 4 bytes per character) it's still a useful comp

  • From TFA:

    âoeIn the last year or two IBM has built machines in the order of a petaflop and in the last couple of weeks IBM announced an ongoing partnership with the US Department of Energy to build a 20 petaflop machine by 2011-2012,â he said.

    âoeWe will need machines which probably have hundreds of thousands of processor cores in them and we roughly know how we can go about engineering it,â he said. âoeIt wouldnâ(TM)t be cost- or technically-feasible to bolt together 50 20-petaf

    • The race is on... (Score:3, Interesting)

      by tjstork (137384)

      Come on... the moment IBM makes a computer with a billion cores, both Microsoft and Linux will be salivating at the change to get -something- to run on them. I mean, what's a GB sized array just to keep track of the CPUs. Pure insanity. Any real geek would love to tackle that.

      • IBM already has 2 operating systems in hand which they have massive experience, AIX and System/Z. They can both run as Hypervisors to Linux so they may use Linux too of course but the real unsurprising thing would be use of ZFS as filesystem for such data. It is already named after zetabyte.

        They also use Plan 9 on Bluegene/L and certainly they aren't doing OS demos on multi million dollar supercomputers, so Plan 9 must be good on some purpose of Bluegene.

        Microsoft? Can they scale really? I mean really, not

    • by Korin43 (881732)
      Yeah just wait for the next release of the Linux kernel: "Now supports over a quadrillion processors and up to 64 yottabytes of memory".
  • From TFA on "racetrack memory":

    Today digital data is stored in two main types of devices, magnetic hard disk drives, and solid state random access memories. The former stores data very cheaply but, since it relies on the mechanical rotation of a disk, is slow and somewhat unreliable. (emphasis mine)

    Define "somewhat". Case in point, I've had a F/W SCSI drive in 24/7/365 operation on my home system for 10 years. Somewhat indeed.

    • Re: (Score:3, Interesting)

      by jandrese (485)
      Anecdotal evidence is the best evidence!

      I have an 800MB HDD that still works, and up until a couple of years ago was in constant use. It was only retired because the old workhorse of a machine it was in was finally replaced. That said, I have also worked with big farms of disks and know that failures happen, and the hard drive is the second least reliable part of most computers after the fans. Anything with moving parts is going to eventually fail, there's no way around it.
    • by kalirion (728907) on Friday September 18, 2009 @05:05PM (#29471517)

      And I have an uncle who smoked a pack a day for 40 years and never got lung cancer.

      • Yes, but what you don't mention is what happened to him in his 41st year. I assume he's 6 feet underground now?

  • ...run Crysis at full resolution!?
  • by MartinSchou (1360093) on Friday September 18, 2009 @05:01PM (#29471487)

    about one thousand million 1Gb memory sticks each day

    First of all, no one would be using manual storage to transfer the data.

    Just throw up some numbers that makes sense to us. Like 99,420.5393 gigabit/second [google.com].

    Most large ISPs use OC-192 [wikipedia.org] as the backbones of their infrastructure. You'd need more than 10,200 [google.com] of those to handle that data load, and that's ignoring the overhead.

    Or to put it into numbers that the RIAA can understand: 1.5707309 * 10^9 [google.com] music CDs every single day.

    At 15 pieces of music per CD and $80,000/song that's $1.88 * 10^15 dollars/day flowing through that network. That's 632 times larger than the US federal budget for 2008 [wikipedia.org].

    No wonder the music industry is in trouble!

  • Imagine a... (Score:3, Interesting)

    by jedigeek (102443) on Friday September 18, 2009 @05:03PM (#29471505) Journal

    Imagine a beowulf of those

    omglol!

  • Oblig. (Score:3, Interesting)

    by drunken_boxer777 (985820) on Friday September 18, 2009 @05:13PM (#29471595)

    Where are the obligatory "beowulfcluster" tags and jokes already?

    Sheesh, the standards around here sure are slipping.

  • Bandwidth (Score:2, Insightful)

    by eccenthink (1312043)
    So if I did my math correctly they're saying if they did distributed computing they'd need to transfer data at a rate of 92.5 Tbps.

    I'm assuming a 1 Gb memory stick is actually 1 GB though...
  • Okay, if I do some rough math, just on the hard drives to dump that to
    assuming 2tb drives, and ignoring the binary/decimal nonsense to be quick
    assuming that the 1eb per day is correct and not the .25eb/day of wiki
    assuming that 2tb costs $100 (volume discount, you know)
    assuming no costs for things to hold these drives, and electricity, etc.

    180 million drives. 18 billion dollars. Per year.

    Let's assume by 2013 we've gone eightfold, to 16tb drives. Good, now we're at 2million ish drives and 2billionish dollars.

    • by nschubach (922175)

      Not all results would have to be saved... only the good ones. (just guessing)

    • by Pretzalzz (577309) on Friday September 18, 2009 @05:45PM (#29471875)
      Maybe they could construct a really fast computer to process the data in real time so they wouldn't have to store it all. They might even release a press release about it.
      • by millia (35740)

        Well, yeah. But it's still got to take a lot of infrastructure that really sounds beyond their capabilities. Let's say:
        16 tb drives, 1 eb= 62,500 drives.
        Let's say 2 weeks of storage; 875,000.
        Let's say 200 per. Let's say $100 extra per for the racks, cases, controllers.
        $262,000,000 just for the storage.

        I dunno, 1.5billion euros, with first usage in 2013 and full capabilities and usage in 2022; maybe they'll make it with that budget if they speak to somebody who already does lots of computation and storage, s

    • by dangitman (862676)

      Okay, if I do some rough math, just on the hard drives to dump that to assuming 2tb drives,

      Where do you get a 2 terabit drive from? (Assuming of course, that a lower-case "t" doesn't mean something other then "tera", which it probably does).

  • IBM is researching an exaflop machine with the processing power of about one billion PCs.

    Or...

    ...a PC with a couple ATI X2 graphics cards in a CrossFire setup.

    But does it run Crysis?

  • My calculator (Qalculate!) tell me, that

    ((1 exabyte) day) ((1 exaflop) second) to byte day
    = 11.574074... (0.1^flop) TB

    That does not make any sense to me. Can someone elaborate? ^^

  • They're gonna need a LOT of pigeons.

  • I lost a pile of dough because of that. Turns out that on a previous race something spooked the horse really badly. Now he never runs well on that track. Of course, the crooked bookie never told me about it.
  • by sker (467551) on Friday September 18, 2009 @05:51PM (#29471915) Homepage Journal

    From the FAQ:

    How far can this telescope see?

    A: ONE STEP BEYOND!

    • Since we are talking about ONE STEP BEYOND, and all that "Madness",

      I propose "This is Ska" by longsy D as the theme song, has that cool
      futuristic acid vibe to it.....

      Youtube link [youtube.com]

    • by smoker2 (750216)
      Come on mods, that was worth more than +3 funny.
      I suddenly had a vision of a load of lab coated techs getting their knees up.
  • by The_Duck271 (1494641) on Friday September 18, 2009 @07:12PM (#29472635)
    Some people might be interested in knowing where all this data comes from. There's a rule of thumb in astronomy that the angular resolution of your images is the wavelength of the radiation you're receive divided by the diameter of your telescope. Radio wavelengths are pretty long (up to tens of meters), so you need really big telescopes, which you get by scattering lots of little telescopes all over the place and then looking at the how the phase of the incoming radiation shifts based on location. So what you do is you sample the voltage of each antenna with 1 or 2 bits of resolution at the Nyquist frequency. So for 100Mhz radio waves you sample at 200MHz. That's 50MB/s for a single antenna. The SKA will likely have tens of thousands of little antennas scattered all over the place. So say 50MB/s times 20,000 antennas = 1TB/s = 100 petabytes/day, which is about what the summary says.

    Now, it's not quite as bad as it looks. You don't have to pipe all this data to a central point to analyze it. You can take a small group of antennas and just look at the correlations between those, combine the data from that group and send the combined data to a second level of correlators, which takes data from a set of small groups, and so on, in a hierarchical fashion. You lose some information this way, but you get most of it, and the only wa to get all the information out of the data would be to bring it all to a central processing location so that data from all antennas could be compared to that of all other antennas, which is O(N^2) in the number of antennas and obviously infeasible for a telescope like the SKA. Even as it is, the system of hierarchical data collection is really pushing Moore's law, as the article shows.
  • by 14erCleaner (745600) <FourteenerCleaner@yahoo.com> on Friday September 18, 2009 @08:20PM (#29473093) Homepage Journal

    the SKA is unlikely to use ... a cloud-based approach

    Well, duh. You can't see anything when it's cloudy.

  • OK, a new unit of measure. Most people that have some idea of what the SKA is could handle a unit like a TB instead of a GB. I suppose they could have used the old 'sheets of paper', or 'Libraries of Books' analogy, so we are lucky in that respect. as for the Billion PCs, are they Xeons? or P-IVs, dual or quad core?

A holding company is a thing where you hand an accomplice the goods while the policeman searches you.

Working...