Several thousand Apple Xserves to drive new EU supercomputer cluster

“A startup on Tuesday unveiled plans at the Apple Expo here to build a distributed supercomputer cluster that could use some thousands of Apple Computer Inc.’s Xserve servers,” Mark Hachman reports for eWeek. “The high-performance network will be installed along Europe’s Internet backbone, startup Omneta plc. said. According to Omneta Chief Technical Officer Tony Clark, the startup, Apple and the European network powerhouse Interoute in October will stage a proof-of-concept cluster involving both Xserve servers and Xserve RAID arrays. If successful, the trial will be extended over the coming year to involve servers placed within Interoute’s facilities, possibly involving as many as 150 data centers across 60 EU cities.”

Hachman reports, “Omneta will sell access to the servers and their processors on a monthly or one-time basis, much as European telcos sell mobile phone service, with either a contract or pay-as-you-go basis. Although the build-out plan is subject to approval and customer demand, upon completion the total network could total ‘several thousand’ Xserves and ‘several hundred’ Xserve RAID arrays, Clark said… According to Clark, Apple representatives said the maker could supply ‘up to 500 servers per day.'”

Hachman reports, “When asked why Omneta chose Apple as a hardware provider, Clark referenced a quote from Apple CEO Steve Jobs, who claimed that the PowerPC G5 was the most cost-efficient, highest-bandwidth processor in the data center. Apple’s Mac OS X will receive a workout, too. ‘According to Apple, theoretically, the Tiger OS is infinitely scalable,’ Clark said. ”Theoretically,’ in plain English, means that they don’t know.'”

Full article here.

Advertisement:
Xserve G5. Enterprise-class features and stability. From $2,999. Free shipping.
Xserve RAID. Up to 7TB of data. From $5,999. Free shipping.

Related MacDailyNews articles:
Australia’s fastest supercomputer to use Apple Xserves, Xgrid and Mac OS X – April 20, 2005
Microsoft’s Windows Server for Supercomputers slips to ‘first half 2006’ – April 06, 2005
Louisville company partners with Apple to create ‘massively fast’ supercomputers for U.S. Navy – December 30, 2004
Virginia Tech’s ‘System X’ Apple Mac supercomputer places seventh in Top500 list – November 08, 2004
Virginia Tech ‘Big Mac’ System X achieves 12.25 teraflops – October 26, 2004
Virgina Tech’s faster ‘Big Mac’ System X supercomputer aims for top-ten world ranking – October 26, 2004
Father of Virginia Tech’s Power Mac G5 supercomputing cluster named among world’s top 100 scientists – September 21, 2004
U.S. Army’s 1,566 64-bit Apple Xserve G5 supercluster can exceed 25 teraflops – September 15, 2004
US Army’s ‘MACH 5’ Apple supercomputer opens eyes – July 05, 2004
US Army’s ‘MACH 5’ Apple supercomputer offers unmatched price/performance – June 24, 2004
COLSA to build ‘MACH 5’ Apple Xserve G5 supercomputer for US Army, to run Mac OS X ‘Panther’ Server – June 21, 2004
‘Big Mac’ Supercomputer may open enterprise market doors for Apple – February 12, 2004
Virginia Tech ‘Big Mac’ supercomputer attracting potential customers – January 28, 2004
Virginia Tech to upgrade ‘Big Mac’ supercomputer to Xserve G5 – January 27, 2004
Virginia Tech Power Mac G5 Supercomputer fires up today – September 24, 2003
Virginia Tech Power Mac G5 Supercomputer costs 90 percent less than nearest competitor – September 23, 2003
Virginia Tech Power Mac G5 supercomputer cluster to run on Mac OS X beta – September 03, 2003
Virginia Tech confirms use of 1,100 Power Mac G5’s for supercomputer cluster – September 02, 2003

16 Comments

  1. So we’ve got superclusters running OS X in the Army, at Virginia Tech, now this one in the EU… The Star Wars special edition DVDs were all image-processed using a cluster of PowerMacs running OS X as well…

    I wonder how many such success stories it will take before people start to realize how totally buff OS X is, and how overengineered it is for home users (in a good way).

  2. This is indeed excellent news and as a citizen of the EU myself I will be writing to my local MEP, Member of the European Parliament, Geoffrey (pronounced Jeffrey) Van Orden with my hopes that the EU will decide to use Apple’s XServes once the concept proves itself worthy, which of course the XServes will.

    Yes I can lobby him if need be too!

    MW: since as in the EU has been anti Microsoft since M$ tried to push for complete monopoly of the EU market and failed and might once again be forced to pay out €300 Billion Euros, $366 Billion US Dollars.

  3. Article quote:
    “Hachman reports, “When asked why Omneta chose Apple as a hardware provider, Clark referenced a quote from Apple CEO Steve Jobs, who claimed that the PowerPC G5 was the most cost-efficient, highest-bandwidth processor in the data center.”
    —–
    Another quote from the linked article “Apple Powers ‘Grassroots’ Supercomputer”:
    The Touring Cluster at the University of Illinois at Urbana-Champaign CSE (Computational Science and Engineering) department “… went with (the G5 based Xserve) … because the price/performance ratio was better than anything else we saw,” Heath said. In addition, he said, CSE had experience working with the G5 processor, also called the PowerPC 970, which CSE had used in products from IBM … he said, “power and cooling was … a critical factor.” That the Xserve G5 produces less heat than a dual-processor server based on chips from Intel Corp. or Advanced Micro Devices Inc. helped the overall cost of the project. “We didn’t want to spend all our money on air conditioning,” Heath said.”

    … So, ahem, excuse my impertinence, but … if this is what all these supercomputer customers are saying, people with a wealth of knowledge regarding which CPUs work best and why … then why is it again that we are moving to x86?

  4. Heh, sort of ironic that Apple, one of the server hardware companies making the *least* noise about “utility computing” now seems poised to power one of the largest deployments of it that I’ve heard about.

    Odyssey:
    “then why is it again that we are moving to x86?”

    Because the move is based on future outlook for both architectures, not current comparisons. PPC is somewhat better right now for power efficiency, etc. But Intel is making a massive shift and pouring their resources in the Pentium M direction; i.e. it’s gonna be all about efficiency instead of raw clock speed. (And it’s about time)

  5. Ryan2004 says: “… the (x86) move is based on future outlook for both architectures, not current comparisons. PPC is somewhat better right now for power efficiency, etc. But Intel is making a massive shift and pouring their resources in the Pentium M direction; i.e. it’s gonna be all about efficiency instead of raw clock speed. (And it’s about time)”

    Thanks for the recap from Apple marketing, but there has been no objective information that’s put out there – by Apple or Intel – that validates that claim. In fact, there have been many documented stories that are examples of just how much Intel is still playing ‘catch up’ in most performance areas, INCLUDING low power technologies. Check out this very recent story:

    “Intel dives into the ultra-low power pool”
    http://news.com.com/Intel+dives+into+the+ultra-low+power+pool/2100-1006_3-5873892.html?tag=cd.top

    After duely noting Intel’s admirable efforts on this front, the story then drops this choice bit of info: “… said Kevin Krewell, a principal analyst at semiconductor research firm In-Stat… “It’s not the first time that this technique has been done, but it’s the first time that Intel has put out a process for these characteristics,” … Texas Instruments and Freescale Semiconductor are among other companies in the chipmaking industry that are working on 65-nanometer transistor leakage issues … TI first introduced … elements of its SmartReflex technologies at the 90-nanometer process node and is now testing the same techniques at the 65-nanometer process level in its family of OMAP 2 chips… Earlier this year, Freescale licensed Virage Logic’s technology for use in Freescale’s upcoming cellular baseband products.”

    So, basically, in this instance Intel is trumpeting the fact that it is implementing a technology that Motorola/Freescale in particular has been using on the G4 for a while now.

    And this is how it is in just about every major tech advance in CPU design that’s of any consequence – from the recent past, as well as projected into the forseeable future. Whether your talking about 64bit, dual core, on chip memory controllers … Intel is either playing second place to AMD or IBM or both, or if they do have a lead (such as in power consumption) it’s in a very narrow market (in this case notebook CPUs) and it’s a tenuous lead at that. In other words, nothing they are doing in the low power area is earthshaking, and in fact every other CPU manufacturer – even lowly Freescale – is close on their heels if not capable of over taking them completely (keep an eye on AMDs upcoming offerings on that front).

    I’m not saying Intel doesn’t have capable CPUs; what I’m saying is that they don’t have the best CPUs out there, nor are they likely too – based on the tech Intel is announcing vs. what their competitors are announcing or already doing – in any future scenario that is in the realm of possibility.

    So, back to my original question – why are we moving to x86 again?

Reader Feedback

This site uses Akismet to reduce spam. Learn how your comment data is processed.