Slashdot Mirror

← Back to Stories (view on slashdot.org)

Virginia Tech on Your Mac Life

Posted by pudge on from the all-you-ever-wanted-to-know dept.

YourMacLife writes "On tonight's Your Mac Life, the Dean of Virginia Tech's College of Engineers, Hassan Aref, will talk about the G5 cluster the college is building and what it means to supercomputing. Questions can be sent in advance to onair@yourmaclife.com." See the web site for more details.

40 comments

  1. Did BSD make this possible? by Not_Wiggins · · Score: 0, Troll

    I checked the article, and all they said was they planned on doing it "more cheaply than buying a supercomputer from a manufacturer."

    Does this mean they're planning on running some variant of BSD? I would imaging that, for licensing sake, they wouldn't put a stock Mac OS on there (OSX)... would cost "too much" and would provide "more" than they need.

    Bravo for the effort... but, methinks they could do this more cheaply (although, not 64-bit) with stock PC hardware.

    --
    Diplomacy is the art of saying, "Nice doggie!" until you can find a rock.
    1. Re:Did BSD make this possible? by coolmacdude · · Score: 3, Interesting

      Reports I've read say they will be running a special version of Darwin designed for high performance clusters.

      --

      -You may license this sig for only $6.99.
    2. Re:Did BSD make this possible? by Juanvaldes · · Score: 3, Informative

      Does this mean they're planning on running some variant of BSD? I would imaging that, for licensing sake, they wouldn't put a stock Mac OS on there (OSX)... would cost "too much" and would provide "more" than they need.
      Unlimited Client X server costs only 1000.

    3. Re:Did BSD make this possible? by Elwood+P+Dowd · · Score: 4, Insightful

      Bravo for the effort... but, methinks they could do this more cheaply (although, not 64-bit) with stock PC hardware.

      Based on the likely purposes of this cluster, that's completely meaningless. This is what 64-bit hardware is for.

      --

      There are no trails. There are no trees out here.
    4. Re:Did BSD make this possible? by AllMightyPaul · · Score: 1

      Wrong. That means that you can have an unlimited number of people connect to your server. It does not mean that you can run an unlimited number of copies.

    5. Re:Did BSD make this possible? by MalleusEBHC · · Score: 1

      Does this mean they're planning on running some variant of BSD? I would imaging that, for licensing sake, they wouldn't put a stock Mac OS on there (OSX)... would cost "too much" and would provide "more" than they need.

      Each Mac comes with a license for OS X. Besides, I'm sure for such a high profile customer, Apple would work out a nice deal for any future upgrades. Even if they didn't, it would only cost ~$75k to buy a new license of OS X for each machine at EDU prices.

    6. Re:Did BSD make this possible? by Danma · · Score: 3, Insightful

      Where were you when this was posted? Clearly Virginia Tech's already done the pricing homework versus commodity PC hardware. Meanwhile, I'm sure Apple will give them a nice deal for buying so many machines (plus free advertising for the ol' fruit company, eh?)

    7. Re:Did BSD make this possible? by dhwebb · · Score: 1

      What does the 10-user licenses cover? I thought if I was using it as a Windows SMB server, that since it was using Samba, it would be already unlimited users because of the GPL license. Does the 10-user count only limit AFP connections or SSH clients or what. Just curious what the user licenses are needed for?

      --
      Only two things are infinite, the universe and human stupidity, and I'm not sure about the former.
    8. Re:Did BSD make this possible? by Daleks · · Score: 4, Insightful

      Bravo for the effort... but, methinks they could do this more cheaply (although, not 64-bit) with stock PC hardware.

      Pundits say the machine is actually cheap. For a 64-bit machine with all the I/O and bus trimmings it is priced nicely. The only thing I'm amazed at is that VT didn't wait for headless cluster-only Xserves. Rack mounting the G5 case looks like it would be a hassle and a shame.

    9. Re:Did BSD make this possible? by The+Bod · · Score: 4, Informative
      Does this mean they're planning on running some variant of BSD?


      Yeah, it's called MacOS X. (early version of Panther)


      I would imaging that, for licensing sake, they wouldn't put a stock Mac OS on there (OSX)


      What? They are buying 1100 machines, they get 1100 copies of MacOS X. What kind of licensing issue are you dreaming of?

    10. Re:Did BSD make this possible? by Anonymous Coward · · Score: 1
      I would imaging that, for licensing sake, they wouldn't put a stock Mac OS on there (OSX)... would cost "too much"


      Huh? Are you implying that they would have to pay more for a Apple with an OS than one without? What cave have you been living in for the past 20 years?


      methinks they could do this more cheaply (although, not 64-bit) with stock PC hardware.


      Maybe you should tell them, because a school with an entire group of people looking at this problem full-time obviously would never have though of checking the prices of PCs. I'm just glad that you're so incredibly smart that you know more about computing clusters than a team of PhDs.


      By the way, OS X is a BSD.

    11. Re:Did BSD make this possible? by ruprechtjones · · Score: 1

      Um, every Mac purchased comes with a copy of OS X. That's 1100 licenses right there. Not sure if each one needs to be the server OS, but I seriously doubt it in a node set-up.

      --
      Kip Hawley is an idiot.
    12. Re:Did BSD make this possible? by Not_Wiggins · · Score: 1

      Point. 8)

      But, it was more of a general comment about the move by Apple to go to a BSD-based OS (compared to OS 9, for instance) that would have made it a viable platform for these advanced uses.

      In retrospect, that point/question wasn't framed as well as I had hoped.

      --
      Diplomacy is the art of saying, "Nice doggie!" until you can find a rock.
    13. Re:Did BSD make this possible? by Senjaz · · Score: 1

      Time is a big factor here, as the article says they've got to finish it by the 1st October for them to enter the super computer rankings for this year.

      So although a cluster of X-Serves would make more sense they can't really wait.

      --
      Don't blame me - this .sig had steal me written all over it.
    14. Re:Did BSD make this possible? by Senjaz · · Score: 1

      Yeah, PC users generally have a hard time understanding that the computer hardware & software are sold as one product by Apple. You're buying a computer that works. No product activation to worry about, no buck passing and finger pointing when you need support.

      There is also no such thing as a "full version" of Mac OS like Windows. All boxed copies are upgrades to whatever you got when you bought it. Mac users tend to forget that when they complain about upgrade prices - why isn't there an ungrade price, why do i have to pay full price? er you are buying the upgrade :P

      --
      Don't blame me - this .sig had steal me written all over it.
    15. Re:Did BSD make this possible? by Anonymous Coward · · Score: 0

      i've noticed that a lot of pc users seem to miss the boat on apple licensing. there is no "pro" version needed, nor do they need to buy the os again via some volume licensing plan if they choose to wipe the machine.

    16. Re:Did BSD make this possible? by RobertEdwards · · Score: 1

      The deadlines for this project procurement are probably related to Federal grant funding cycles. The US Government fiscal year kicks over in October.

    17. Re:Did BSD make this possible? by The+Bod · · Score: 1

      They probably didn't consider this, but resale could be a factor as well. When they upgrade their Super Computer to a G6 in a few years, I would guess that they would have an easier time reselling a bunch of PowerMacs than a bunch of Xserves. How cool would it be to say you have a piece of a Super Computer?

    18. Re:Did BSD make this possible? by switcha · · Score: 1
      ...I'm amazed at is that VT didn't wait for headless cluster-only Xserves.

      I believe they were in a rush to beat the cutoff date for the Linpack Top 500 Supercomputer List

      Apple gave them the performance and, I'm sure, the sweet deal they needed to make it happen.

      --
      You know what? ... A little club soda *did* get that out!
    19. Re:Did BSD make this possible? by djward · · Score: 1

      Actually, they paid regular .edu institution prices for all machines. They were ordered through the Apple store online. Apple did give them a break on RAM - Crucial RAM at wholesale (NOT Apple) prices, preinstalled before the machines shipped.

    20. Re:Did BSD make this possible? by n8_f · · Score: 1

      Nope, the Dean of the Engineering school says they are using Mac OS X for now. They might look at something else in the future, but for now it is 10.2.7 (the only OS that fully supports Apple's G5).

  2. Why the short time frame? by Anonymous Coward · · Score: 0, Interesting

    According to comments on previous stories, the main factor was delivery time. Could you comment on this? As well, why is it important to rush in order to make the next top computing list?

    1. Re:Why the short time frame? by digital_freedom · · Score: 1

      They are trying to make the deadline for the Top 500 list of supercomputers here. The conference is in early November and they need to be able to run the LinPack benchmarks on the system in order to get a spot. They will hopefully make it into the Top 5.

  3. Can I Play Doom III On It? by Anonymous Coward · · Score: 2, Funny

    Well, can I?

  4. Re: Here's the clue ;-) by bursch-X · · Score: 2

    Wrong. That means you can have an unlimited number of people connecting via AFP.

    The 10 user license only restricts the number of people that can simultaneously connect via AFP, there are no restrictions on the number of users created or the number of users simultaneously connecting via SAMBA, FTP or NFS.

    --
    There are two rules for success:
    1. Never tell everything you know.
  5. Re: Here's the clue ;-) by AllMightyPaul · · Score: 1

    Eh, I didn't indicate a protocol, but that doesn't make my original point any less valid and that was that you can only install it once even with the unlimited client access license.

  6. Re: Here's the clue ;-) by dhwebb · · Score: 1

    OK. You answered my previous question about wht the licenses were for. Now another question, why would you buy AFP licenses if obviously the Macs can talk over SMB or NFS for free? What is the AFP advantage that makes it worth paying for? I'm sure Samba is more difficult to configure that AFP but if your running a server, you should have the ability to handle the few extra steps.

    Once again, just curious?

    --
    Only two things are infinite, the universe and human stupidity, and I'm not sure about the former.
  7. $5m??! by mantera · · Score: 0, Troll

    isn't 5m a lot of money for a top ten supercomputer? I thought you could get cluster computing on the cheap. Would they have had it cheaper had they used linux and intel?

    1. Re:$5m??! by Clock+Nova · · Score: 1

      According to the many articles I've read concerning this, they compared price and performance with a lot of different companies and machines. So, no. They wouldn't have gotten it cheaper with Linux and Intel.

      --
      There they were, sitting in the van with all those dials, and the cat was dead. -V. Marchetti, CIA
    2. Re:$5m??! by adam872 · · Score: 2, Informative

      Compared to the ASCI series put together by the US government, $5M is not a lot of money. Consider that a single Sun SF15k, IBM p or z series or HP SuperDome can cost easily this much (that's just one machine).

      I think VT are getting a pretty good deal. Really large clusters/supercomputers *can* cost upwards of $30M, depending on the configuration.

    3. Re:$5m??! by Anonymous Coward · · Score: 0

      Compared to the ASCI series put together by the US government, $5M is not a lot of money. Consider that a single Sun SF15k, IBM p or z series or HP SuperDome can cost easily this much (that's just one machine).

      Yes, but those aren't comparable. The IBM supercomputers have insanely fast networking (IBM 'switch2' in a hypercube configuration, with dedicated communication CPUs), and machines like the superdome are single-image machines.

      A single machine with e.g. 64 CPUs sharing memory is orders of magnitude more expensive than 64 desktop PCs, for a reason.

  8. Re: Here's the clue ;-) by justforaday · · Score: 1

    if you set up the machines to use remote home directories from an os x server it will use AFP by default. also, i believe in earlier versions of os x, the server browser only showed AFP servers. i could be wrong on that last point tho. either way, they've certainly made the os play a lot nicer with various connection methods over the years.

    --
    I'll turn into a supernova and burn up everything. Well I'll turn into a black little hole and you'll turn into string.
  9. Notes From the Interview by n8_f · · Score: 2, Informative
    Well, I browsed through the links at threshold=1 and I couldn't find any comments from anyone who had actually listened to the broadcast in question. No surprise there, I guess.

    It wasn't that great, so you didn't miss much. It starts at 1:17 and ends at about 1:37 in the archive file for those who would like to listen. For everyone else, here is a list of the highlights:
    They didn't go with an Xserve chassis because they couldn't get one with a G5.

    They are using the stock chassis, no modifications.

    It will take up about 6 to 8 racks. (Which seems really small to me; they must be packing them in tight).

    Using OS X for now, the stock install (10.2.7). They have had other people approach them and they will consider other OSes in the future, but they are going with an OS X install for now (this seems to be a lack-of-support issue).

    They do have a 400-node cluster running now, so they aren't complete novices to building clusters. But this was still very new for them and several times Mr. Aref mentioned that Apple had helped them out a lot, going to some of the other vendors involved and talking with them on VT's behalf, etc. Apple obviously sees this as a very strategic project for them.

    They chose 1100 nodes because that hit the sweet spot for budgeting and where they wanted to be on the TOP500 list.

    The cluster comprises over 19 tons of gear.

    It will be a node on a network of supercomputers nation-wide - the National Lambda Rail initiative.

    They (VT) are creating a video documentary of the whole project that it sounds like they will put online later in the year. They are also writing a handbook as a kind of howto for building your own cluster. Sound very willing to talk to other groups about how they did it and help others along.
    That is about it. Not much as far as technical details. Mr. Aref said they will release all of the technical details later. He wouldn't venture a guess on where they will be on TOP500 until he's seen some benchmarks, but they obviously expect to make the top 10.
    Personally, I am extremely curious about this whole project. Using a desktop chassis seems like an unconvential way to build a large cluster, so the photos will be very interesting. But the documentary! I think that is awesome. This might provide a unique perspective into how a large cluster is built (Mr. Aref joked that it involved a lot of pizza).
    1. Re:Notes From the Interview by Anonymous Coward · · Score: 0

      It will take up about 6 to 8 racks. (Which seems really small to me; they must be packing them in tight).

      That must be a misunderstanding - he must have meant 6-8 dual machines per rack.

      A rack is 43U, so even without accounting for switches or stuff you could only pack 344 Xserves in 8 racks. There is no way this is going to be 4 times higher density with desktop cases.

    2. Re:Notes From the Interview by n8_f · · Score: 1

      I cut out the relevant question and put it up here, so you can judge for yourself, but I think the context and quote are pretty clear. In referring to the VT datacenter, he says: "They've configured one corner of that - umm - I think there are about 6 to 8 racks or maybe more with all of these machines standing on them." He may be somewhat removed from the minute details of the project, but if he was off by over a full magnitude of order, why would he throw out any number? And he is definitely saying "6 ta 8", not "6ti 8", "6 ti 8", or even "6ta 8".
      BTW, in the archive files, the relevant quote is at 1:23:55, with the question starting at 1:23:40.

      A rack is 43U.
      A standard rack is 43U. Just after the quote above, at about 1:25, he mentions that Liebert custom built racks for them. And given that they are accommodating desktop chassis, I doubt they are using a design based on a standard rack (especially with regards to footprint).

      But I guess we'll see when the pictures come up. I must admit, it was one of the odder things he said and I had to listen to it several times to make sure I wasn't misunderstanding. It has my curiosity piqued.

  10. DOA machines by djward · · Score: 1

    I just heard through one of the people involved that 900 machines have arrived now. Out of those, only 2 were DOA. That's less than 0.25%. Impressive. And only 200 more to go.

  11. I can tell you one way it's affecting my mac life by fuzzeli · · Score: 1

    having to wait another month to get my hands on my preordered dualie! I was so upset, I bought an iMac for my wife, and now I feel much better.

    I don't know what kind of subliminal messages Steve puts in his keynotes, but my wallet opens itself around apple logos these days. Scary.