University of Illinois preps 640-node Apple Xserve G5 cluster

Posted:
in Future Apple Hardware edited January 2014
The University of Illinois is home to yet another Mac supercomputing cluster, AppleInsider learned.



Located on the University's main campus at Urbana-Champaign, the new Turing Cluster currently consists of 640 dual-processor Xserve G5s, each with two G5 processors running at 2GHz and 4GB of RAM, for a total of 1280 processors. A high-bandwidth, low-latency Myrinet network from Myricom is being used to connect the cluster machines.



The cluster uses wired Ethernet networking routed by full-duplex 100Mb switches from Cisco Systems. It also features a 1Gb link between the front-end array and the primary Cisco switch, and runs a version of Apple's Mac OS X Server 10.3 "Panther" operating system.



Until recently, the University's Turing Cluster ran version 7.2 of Red Hat Linux using only 208 dual-processor machines, which were supplied by both Dell and Hewlett-Packard. But last summer, Michael Heath, Director of Computational Science and Engineering, saw that the Intel-based cluster had reached the end of its life span in terms of reliability. Additionally, the cluster was no longer generating interest with its aging computational ability, he said.



Heath set out on a mission to revamp the Turing Cluster into a facility that would be open to local users, scientists, and the University's student population of over 35,000. He presented his endeavor to several corporations, of which, he said, Apple responded quite eagerly.



After generating the necessary funding through donations from over a half-dozen University departments -- including the Beckman Institute, College of Engineering, and Department of Computer Science -- Heath and a team of supporters began nailing down the details of the cluster with Apple. By October 2004, they had constructed a 64-node mini-cluster in a temporary server room as a small-scale model of the entire system.



Speaking with AppleInsider this week, Heath revealed that the full-scale 640-node Xserve G5 cluster has been fully functional for about a month now, but is not yet sufficiently stable. He said it would take approximately one more month to stabilize the cluster, acquire a few missing parts, and exchange a couple of Xserve G5 units, which were dead-on-arrival.



The new Turing Cluster from Start to Finish Click images for larger view.



The cluster is housed in a newly renovated server room that can handle a cooling load of up to 550,000 BTU per hour and supports over 45 tons of cooling capacity using four distinct cooling systems -- three of which can adequately support the cluster at any given time.



In the near future, the University hopes to be able to perform tests to rank the cluster's computational ability, which, according to its specifications, could peak at nearly 10 teraflops.



Heath declined to provide specifics of the University's arrangement with Apple, but a web page dedicated to the new Turing Cluster says Apple provided the hardware under a combination purchase/donation agreement.



And while Health conceded that there is no formal upgrade plan for the future of the cluster, he said his team has been seriously considering expanding the cluster to 1280 nodes, double its current size.
«1

Comments

  • Reply 1 of 21
    nagrommenagromme Posts: 2,834member
    Cool. FYI, I've been keeping a list of Mac supercomputer bookmarks. Xserve clusters seem to be big lately. Price/performance, reduced heat output, and UNIX with easy management are often cited as reasons.



    There are now more than a dozen Mac OS X / Xserve clusters of 32 processors or more--eight with 200 processors more, four with 1000 processors or more--topping out at 3,132 for MACH 5:



    * 1280 processors - U. of Illinois' "Turing Cluster"

    (640 dual 2.0 nodes, possibly to be doubled, used for a range of academic research and replacing a Dell/HP Linux cluster)



    * 250 processors - U. Pitt's Human Genetics cluster

    (125 dual G5 nodes used for genetics research)



    * 200 processors - GeoCenter cluster

    (100 dual 2.0 nodes used for seismic data processing)



    * 1344 processors - French CGG cluster

    (672 dual nodes, integrated into an existing 40 TFLOP cluster for oil prospecting)



    * 48 processors - Louisiana State's "Nemaux"

    (24 dual G5 nodes with Xgrid, used for 3D animation, audio, and scientific computing)



    * 2200 processors - VA Tech's "System X" aka "Big Mac"

    (1100 dual 2.3 nodes and counting, used for a range of academic research)



    * 256 processors - UCLA's "Dawson"

    (128 dual 2.0 nodes, used for plasma physics research)



    * 32 processors - Australian Defence Force's "Checkmate"

    (16 dual nodes, used for command and control simulations)



    * 86 processors - UNC's cluster

    (43 dual nodes, used for proteomics research)



    * 76 processors - UC Davis's cluster

    (38 dual nodes, used for Genome Center research)



    * 72 processors - UC Santa Cruz's cluster

    (36 dual nodes and counting, used for a range of academic research)



    * 3132 processors - US Army's "MACH 5"

    (1566 dual nodes, used by the Army and NASA for hypersonic flight research)



    * 512 processors - U. Maine's "Baby MACH 5"

    (256 dual nodes, used for software development and optimization for MACH 5)



    Also, the US Navy is using Xserves on board submarines, to run their Linux-based sonar imaging system.



    And VA Tech is considering much larger Mac clusters in future: "System L" and "System C."
  • Reply 2 of 21
    placeboplacebo Posts: 5,767member
    Whoa, go Apple!
  • Reply 3 of 21
    screedscreed Posts: 1,077member
    Doing the math... 9488+ G5s!



    Screed
  • Reply 4 of 21
    gene cleangene clean Posts: 3,481member
    Hey, thats my college!!!
  • Reply 5 of 21
    nagrommenagromme Posts: 2,834member
    9,488 G5 chips... and IBM has only made 10,000!



    It all makes sense now...
  • Reply 6 of 21
    This is going to be fun in large companies when the R&D guys put in a request for a Mac supercomputer and the IT guys tell them they should buy a Dull. Dull won't have a chance and the CEO will be asking IT why they can't start looking at the Mac mini for a lot of replacement orders -and the IT guys ask "What's a Mac mini?" The CEO will refer them to his 15 year old son . . .
  • Reply 7 of 21
    kennethkenneth Posts: 832member
    Alright, one of my friend goes to U of I, Urbana-Champaign.
  • Reply 8 of 21
    As we all know, HAL was born at Uof I. HAL's name was based on one letter to the left of IBM I-H, B-A, M-L. But, what if it turns out that HAL is really a Mac with an IBM CPU?



    WHAT DOES IT ALL MEAN?
  • Reply 9 of 21
    Wait, first it says it's Myrinet connected, then it says it's ethernet connected. So it's both?
  • Reply 10 of 21
    dcqdcq Posts: 349member
    Quote:

    Originally posted by nagromme

    Cool. FYI, I've been keeping a list of Mac supercomputer bookmarks. Xserve clusters seem to be big lately. Price/performance, reduced heat output, and UNIX with easy management are often cited as reasons.



    There are now more than a dozen Mac OS X / Xserve clusters of 32 processors or more--eight with 200 processors more, four with 1000 processors or more--topping out at 3,132 for MACH 5:



    [snip]



    And VA Tech is considering much larger Mac clusters in future: "System L" and "System C."




    Dude: Think about it. If VATech spent $5.8 million on Big Mac with 17+teraflops...Apple should spend $50 million of its iPod windfall, build a 100+teraflop supercomp, donate it to some worthy cancer/AIDS/medical-whatever research institute. They'd get crazy feel-good press, a mega tax-write off, and they'd get instant bragging rights for building the fastest computer on the planet for the equivalent of pocket change. Three times faster than the Earth Simulator for 1/8th the cost!! And they'd probably make back the $50 million in a matter of hours as universities, corporations, and governments (not to mention assorted international criminal conspiracies and mad-scientists bent on world domination) scrambled to gobble them up.
  • Reply 11 of 21
    slugheadslughead Posts: 1,169member
    Quote:

    Originally posted by DCQ

    Dude: Think about it. If VATech spent $5.8 million on Big Mac with 17+teraflops...Apple should spend $50 million of its iPod windfall, build a 100+teraflop supercomp, donate it to some worthy cancer/AIDS/medical-whatever research institute. They'd get crazy feel-good press, a mega tax-write off, and they'd get instant bragging rights for building the fastest computer on the planet for the equivalent of pocket change. Three times faster than the Earth Simulator for 1/8th the cost!! And they'd probably make back the $50 million in a matter of hours as universities, corporations, and governments (not to mention assorted international criminal conspiracies and mad-scientists bent on world domination) scrambled to gobble them up.



    And whoever they donated to would be bankrupt once the first power bill and rent came in... and then there's the labor, software writing, and upkeep.



    We're only looking at the fixed costs here, I bet VATech spends 700k/yr on upkeep, and probably 300k initial setup cost (not including the equipment).



    You're also forgetting that apple's net income pre taxes was $276m for last quarter.. dumping 50m from that into some PR stunt would piss of stockholders something awful.
  • Reply 12 of 21
    macroninmacronin Posts: 1,174member
    Quote:

    Originally posted by slughead

    And whoever they donated to would be bankrupt once the first power bill and rent came in... and then there's the labor, software writing, and upkeep.



    We're only looking at the fixed costs here, I bet VATech spends 700k/yr on upkeep, and probably 300k initial setup cost (not including the equipment).



    You're also forgetting that apple's net income pre taxes was $276m for last quarter.. dumping 50m from that into some PR stunt would piss of stockholders something awful.




    Or they could NOT donate the cluster to anyone and hang on to it themselves...



    Sell time to gov/corp/mil users; donate job runs to non-prof/itcharity/medical causes...



    The former will pay the bills, the latter will generate regular positive PR...



    And the Mac geeks would have a new Mecca...
  • Reply 13 of 21
    sunilramansunilraman Posts: 8,133member
    all this raw, extreme G5 processing power makes me want to..... *passes out*



  • Reply 14 of 21
    What a coincidence, I started reading Heath's book "Scientific Computing: An introductory survey" today....
  • Reply 15 of 21
    wmfwmf Posts: 1,164member
    Quote:

    Originally posted by ClimbingTheLog

    Wait, first it says it's Myrinet connected, then it says it's ethernet connected. So it's both?



    Yes. The two networks are used for different kinds of traffic; usually MPI goes over Myrinet and everything else goes over Ethernet.
  • Reply 16 of 21
    g3prog3pro Posts: 669member
    Quote:

    Originally posted by Gene Clean

    Hey, thats my college!!!



    What are you studying?





    Here's a link to the homepage:



    http://www.cse.uiuc.edu/turing/

    http://www.cse.uiuc.edu/turing/announce.html

    http://www.cse.uiuc.edu/turing/gallery.html

    http://netfiles.uiuc.edu/benoc/clust...r%20album.html
  • Reply 17 of 21
    sunilramansunilraman Posts: 8,133member


    hey on one of the photo gallery there's just something called "Glowing Red Button" Mysterious!!!







    "hit this button in case of G5 cluster developing its own self identity and starts implementing plans for world domination"
  • Reply 18 of 21
    Quote:

    Originally posted by jonboy

    As we all know, HAL was born at Uof I. HAL's name was based on one letter to the left of IBM I-H, B-A, M-L. But, what if it turns out that HAL is really a Mac with an IBM CPU?



    WHAT DOES IT ALL MEAN?




    Arthur C Clarke has always strenously denied that HAL was IBM one letter removed - it's a coincidence!
  • Reply 19 of 21
    wmfwmf Posts: 1,164member
    The Big Red Button is probably an emergency power off. It's intended to be pressed in case of a fire.
  • Reply 20 of 21
    sunilramansunilraman Posts: 8,133member
    Quote:

    Originally posted by wmf

    The Big Red Button is probably an emergency power off. It's intended to be pressed in case of a fire.



    Is everything alright Dave? Dave, please don't go near the big red button...



    Dave, the temperature in the room is within agreed parameters, and I do not detect a fire. Dave, you're not thinking of turning my G5 cluster off, are you, Dave?



    Dave, I've noticed that you've been breathing a bit heavier and your pulse goes up everytime you're in the cluster room... Is everything alright, Dave?




Sign In or Register to comment.