Forgot your password?
typodupeerror
Intel Technology

Intel Unveils 6-Core Xeon 7400 235

Posted by kdawson
from the not-cheaper-by-the-half-dozen dept.
JagsLive recommends CNet coverage that begins "Intel officially unveiled its six-core 'Dunnington' Xeon 7400 processor Monday ... As expected, Intel launched the Dunnington chip for high-end servers ... The Xeon 7400 is also one of the first Intel chips to have a monolithic design. In other words, all six cores will be on one piece of silicon. To date, for any processor having more than two cores, Intel has put two separate pieces of silicon ... inside one chip package."
This discussion has been archived. No new comments can be posted.

Intel Unveils 6-Core Xeon 7400

Comments Filter:
  • by Yvan256 (722131) on Tuesday September 16, 2008 @08:11AM (#25023573) Homepage Journal

    I'm betting new Mac Pros will be launched today.

  • Base 2 (Score:4, Interesting)

    by daveime (1253762) on Tuesday September 16, 2008 @08:12AM (#25023581)

    Is it just me, or does 6 seem like a counter intuitive number of cores ?

    2,4,8,16 ... we've been using binary since the start, now we have to start in trinary ?

    • Re:Base 2 (Score:5, Funny)

      by Anonymous Coward on Tuesday September 16, 2008 @08:21AM (#25023651)
      6 = 8 - 2 broken cores ?
      • by DrYak (748999) on Tuesday September 16, 2008 @09:33AM (#25024357) Homepage

        6 = 8 - 2 broken cores ?

        You joke but that's already the case with PS3's Cell (7 SPU = 8 - 1 broken), with tripple core Phenom (3 = 4 - 1 broken), and with a very high number of graphic cards (The range segment {pro/mid/low-cost} on which a GPU is used = the number of functional cores they managed to salvage)

        A separate reason may be the number of {quickpath/hypertransport/etc.} interconnects (6 cores require 15 interconnect to communicate, 8 cores require 28 interconnects). 6 to 8 cores isn't such a big increase but keeps the number of inter connect reasonnable.
        (Other processors types like Tilera end up only interconnecting adjacing cores on their 64x chips and you have a strongly *Non*-Uniform Architecture, with not all core able to reach and talk to others at the same speed)

        • by jkenneth24 (962795) on Tuesday September 16, 2008 @10:36AM (#25025243)
          i dont think this uses Quick Path Interconnects yet... the article stated this was still Penryn. Theres also a bit at the end where the articles stated AMD chiming in and saying "Intel has taken the old front-side bus architecture and added 6 cores to it,"
        • by Kjella (173770)

          You joke but that's already the case with PS3's Cell (7 SPU = 8 - 1 broken), with tripple core Phenom (3 = 4 - 1 broken), and with a very high number of graphic cards (The range segment {pro/mid/low-cost} on which a GPU is used = the number of functional cores they managed to salvage)

          Just because they're disabled doesn't mean they're broken. Say you have a 95% yield on a PS3 core. 8 SPU good = 0.95^8 = 66%, 7+ SPU good = 0.95^8+8*0.95^7*0.05 = 94%. Yet they can't use the full power of the 66% perfectly fine processors without creating two console versions. In other things like CPUs and GPUs they do go in different bins of functional cores/lines/clockspeed but it also depends on what the market wants so they can be downbinned and sold as cheaper parts. In the past people have been buying

    • Re: (Score:3, Informative)

      by houghi (78078)

      There is no need for a base two if you are adding. You add cores, you do not multiply them.

      2, 3 (yes, triple code do exist), 4, 6. I guess the next step will be 8 and perhaps even skip that and make it 9 (3x3).

      • by telchine (719345)

        3 (yes, triple code do exist)

        The only three core processors [theinquirer.net] I know of are effectively defective quad core processors.

        The 4th core is defective so, rather than disposing of them, they are sold as tri-core processors.

        Whilst there's no requirement for base two, there is usually a requirement for an even number of cores in SMP (symmetric multi processing)

        • Re: (Score:2, Informative)

          by adrianwn (1262452)
          The "symmetric" in SMP refers to all CPUs being identical, not to the actual number of processors.
          • by telchine (719345)

            The "symmetric" in SMP refers to all CPUs being identical, not to the actual number of processors.

            I stand corrected. Thanks

        • Re: (Score:2, Informative)

          by dumael (1172411)
          Symmetric multi-processing refers either the generally the type of cpus, i.e. all processors have the same capabilities, or to their relationship with memory. SMP is generally shared memory systems with a set of uniform processors.
          http://en.wikipedia.org/wiki/Symmetric_multiprocessing [wikipedia.org]
        • Re:Base 2 (Score:4, Informative)

          by mr_mischief (456295) on Tuesday September 16, 2008 @10:03AM (#25024759) Journal

          That has nothing to do with symmetric multiprocessing. SMP means that all the chips can make memory accesses to all the memory at the same speed. It is the opposite of Non-Uniform Memory Access, or NUMA, in which certain processors (some or all of them) take longer to talk to certain parts of main memory than others or systems in which processors have a faster path to some private off-chip memory in addition to the main shared memory.

      • Again, this is a defective four core chip which you've been duped into purchasing.

        • by mr_mischief (456295) on Tuesday September 16, 2008 @10:05AM (#25024777) Journal

          That's no different from the 486SXes, many of which were 486DX parts with the defective math coprocessor diked out. It's not very different from how the clock rate on every mainstream chip is determined by how many chips turn out to be stable at which speeds.

        • by confused one (671304) on Tuesday September 16, 2008 @10:12AM (#25024893)
          What's wrong with that. Intel has sold defective processors for years, either binning them at a lower clock speed or trimming out chunks of cache and selling them as Celerons. If it can serve a purpose, does the job you need to get done, and it's available cheap... I don't see a problem here.
        • by MightyYar (622222)

          Again, this is a defective four core chip which you've been duped into purchasing.

          In the industry, we call it a "reworked" part. It's not being "duped" if it costs less and does what it advertises.

          By the way, your memory is almost certainly "defective" by your definition. Memory is reworked like crazy - loads of extra traces. And if it's REALLY bad, the top-tier guys sell it to discounters to sell as lower-capacity part.

        • You're really being sold 3 tested and working cores. You're supposed to ignore failed core on the die. It has nothing to do with the 3 that do work.

    • Re: (Score:2, Informative)

      by zeridon (846747)

      see amd ... they have 3 cores in one chip and the shit fares prety well

      • by Ngarrang (1023425)

        see amd ... they have 3 cores in one chip and the shit fares prety well

        AMD once said there was a greater efficiency in interconnecting 3 cores, compared to 4.

        • Re:Base 2 (Score:4, Interesting)

          by dubbreak (623656) on Tuesday September 16, 2008 @11:17AM (#25025807)

          AMD once said there was a greater efficiency in interconnecting 3 cores, compared to 4.

          Anyone with simple graph theory can see that. 3 nodes only need 3 edges(or interconnects in this case) so that they are only one hop away from each other. With 4 nodes you need 6 edges for a complete graph / clique (each node is adjacent). Now whether that is applicable to the amd chip is a whole other question. I don't have the time (nor do I care to) look up the implementation of the chips.

      • by Spatial (1235392)
        Faring pretty well my arse. I still don't have a worthwhile upgrade to my X2 6000 and it's almost two years old. Right now all the AMD CPUs available are downgrades or trade-offs. I want an upgrade. Even the priciest, most powerful Phenom they have to offer is slower than the X2 6000 in every application except ones which fully utilise three or more cores. (almost none)
    • Re:Base 2 (Score:5, Funny)

      by davidbrit2 (775091) on Tuesday September 16, 2008 @08:24AM (#25023679) Homepage
      Hey, 6 is a power of 2. It's 2^2.585, to be inexact.
    • Re: (Score:3, Informative)

      Is it just me, or does 6 seem like a counter intuitive number of cores ?

      Remember they need to put other stuff on the silicon too. The XBox 360's CPU uses three quarters of the die for three processors and puts the shared cache etc. in the fourth quarter. Six + support circuitry probably fits a square die better than eight + support.

      • Re: (Score:3, Informative)

        by blind biker (1066130)

        Sun had 8-core server CPUs since several years now. They didn't have any problems allocating the die's surface.

    • by oldhack (1037484)

      Is it just me, or does 6 seem like a counter intuitive number of cores ?

      Yeah, especially with the new motherboard that fits three of these together...

    • we've been using binary since the start, now we have to start in trinary ?

      You need to be more flexible. I seem to recall from my Number Theory class that the base you choose can be somewhat arbitrary. In fact, I've heard that a lot of people use a Base 10 system!

      The AMD Phenom comes in a 3 core variety (which may or may not be a "broken" 4 core chip). The Xbox 360 has a 3 core processor. The cell processor used in the Sony Playstation 3 and IBM blade servers is a 9 core (granted only 1 is a full core). There's no good reason for limiting yourself to powers of 2...

      • by jacquesm (154384)

        Sperry Univac actually had a machine with 9 bit bytes... those were the days ! Just think about what you could do with that extra bit...

      • by hedwards (940851)

        You need to be more flexible. I seem to recall from my Number Theory class that the base you choose can be somewhat arbitrary. In fact, I've heard that a lot of people use a Base 10 system!

        That would be correct, the main issue with some bases like binary is that they're difficult for humans to read and take up a huge amount of space on paper. There isn't any inherent reason why we could be using hexadecimal or larger base.

  • Specs? (Score:4, Insightful)

    by sanosuke001 (640243) on Tuesday September 16, 2008 @08:17AM (#25023615)
    There wasn't much in terms of technical specs in TFA. 6 cores, 16MB cache, anything else? Clock speed? 16MB of L2? L3? FSB? DDR(n)? (Though this is probably more up to the MB manufacturer) Why are they moving the memory controller off silicon? That in itself seems like a step backwards.

    I would like to see them pushing consumer multi-core computing more personally. Get MS and other application manufacturers to support more cores. Servers have been doing it for ages and with pretty much all consumer level chips being dual core they should be pushing this angle more.

    Though, them incorporating all of the cores on a single piece of silicon is definitely a step forward; the lack of additional specs and the notion of moving the memory controller make this seem like not as big of an announcement...
    • Re:Specs? (Score:5, Insightful)

      by morgan_greywolf (835522) on Tuesday September 16, 2008 @08:25AM (#25023701) Homepage Journal

      I would like to see them pushing consumer multi-core computing more personally. Get MS and other application manufacturers to support more cores. Servers have been doing it for ages and with pretty much all consumer level chips being dual core they should be pushing this angle more.

      And before anyone says...."yeah, but Linux/Mac OS X supports multi-cores out of the box".... Yes, yes it does. However, most of the applications don't actually benefit much from SMP by themselves. A few things like video conversion, but, for the most part, office suites, e-mail user agents, etc., do not actually benefit directly from SMP.

      OTOH, why should they? Any processor made within the last five years is good enough for that stuff.

      • Re: (Score:3, Funny)

        OFC if your dealing with server, then it does make a difference for example Unisys 96 core offering (*nix & possible mac only) would be able to hold 256 SQL server databases ... or vista pro

      • by Macrat (638047)
        Wow, if you need your e-mail and office suite to be processor aware, maybe you should stop using Windows.
      • You can, on most operating systems, actually run more than one application at a time now.

    • Re: (Score:3, Informative)

      by billcopc (196330)

      There wasn't much in terms of technical specs in TFA. 6 cores, 16MB cache, anything else? Clock speed? 16MB of L2? L3? FSB? DDR(n)? (Though this is probably more up to the MB manufacturer) Why are they moving the memory controller off silicon? That in itself seems like a step backwards.

      Pentiums have never had an on-die MMU in the first place, they're actually doing that for the upcoming Nehalem. AMD has used an on-die MMU since the Athlon 64, and while it certainly helps squeeze more efficiency out of the system, the faster clocked bus on Intel rigs often made up for the theoretical performance gap.

      The Clock speed was omitted from the article, but the Xeon 7400 is clocked at 2.6ghz. That's actually pretty decent for a Xeon, they don't go as high as the mainstream processors because they

  • by halcyon1234 (834388) <halcyon1234@hotmail.com> on Tuesday September 16, 2008 @08:22AM (#25023663) Journal
    ... and a moisturizer strip for a cool, refreshing finish.
  • Wattage (Score:5, Interesting)

    by locster (1140121) on Tuesday September 16, 2008 @08:24AM (#25023685)

    I think server builders these days are less interested in the number of cores per CPU and more interested in improvements in the performance/wattage ratio.

    • Watt you say?
    • by Kjella (173770)

      That depends entirely where those servers are. Many places rack space comes at a huge premium and anything packing more punch into less space is valuable. Of course, if you're talking about your local server room where you could just put in another rack if you needed to it's not that big a deal.

      • Re: (Score:3, Informative)

        by afidel (530433)
        Any decent host is also going to kill you for bringing more power to your rack so there is definitely a balancing point. The fact is that the datacenter is designed for a certain power density and going beyond that really screws things up. Airflow and cooling densities, percentage of space allocated to UPS and generators, etc. But, looking at the Intel press release, these suckers pack 6 cores into a 65W power envelope, quite impressive. This compares quite favorably with the 45W for the best current genera
    • Well, considering this 45nm part uses less power for 6 cores than the 65nm parts from the same company use for 4 cores, I'd call that win-win.

  • by nimbius (983462) on Tuesday September 16, 2008 @08:24AM (#25023689) Homepage
    i might finally be able to play crysis on my vista ultimate machine? i mean, granted, my pc will look more like a LHC when im through with it...but a few black holes are worth it
  • With a catch.... (Score:4, Informative)

    by Ritz_Just_Ritz (883997) on Tuesday September 16, 2008 @08:25AM (#25023699)

    "There's an odd catch, however, that will affect the highest of high-end configurations. "Because Microsoft Windows operating system support is limited to a 64-core environment, within a single OS instance, we'll support up to 64 cores," said Colin Lacey, a Unisys marketing vice president."

    Gads, who on earth would run a 64-core Windows box? Unless they want to virtualize out multiple servers on one bit of hardware. Most of the "heavy lifting" I've seen on servers with mucho processor cores are running some flavor of Unix. I'm kinda surprised this hasn't been fixed already given the momentum of multi-core processors.

    Cheers,

    • Re:With a catch.... (Score:5, Interesting)

      by ShadowRangerRIT (1301549) on Tuesday September 16, 2008 @08:31AM (#25023749)
      Windows optimizes for the low core case. I believe they use a bit field to keep track of the cores, so the 32 bit flavors of Windows are limited to 32 cores, while the 64 bit versions are limited to 64 cores. There may be a high end server SKU that bypasses that limitation, but I don't know of it.
      • by afidel (530433)
        There isn't, if Unisys doesn't have it no one does. They basically are billed as the mainframe maker of the x86 market. They were the ones that spurred the development of the Windows Datacenter edition development. There are some VERY large databases running on Unisys hardware.
      • by Kingrames (858416) on Tuesday September 16, 2008 @09:36AM (#25024391)
        64 cores should be enough for anybody.
      • SKU? Do you work in retail?

        At the moment,
        Windows Datacenter [microsoft.com] supports 32 cores on X86, and 64 cores on x64 and Itanium.

        The implication of "SKU" is that this limitation is trivial, and is imposed purely by marketing considerations.

      • by PeeAitchPee (712652) on Tuesday September 16, 2008 @10:08AM (#25024817)

        Gads, who on earth would run a 64-core Windows box?

        The ABBYY OCR engine (Windows only) in any of its latest versions (either direct from ABBYY or OEM'd into someone else's product) will multithread during recognition -- one thread for each core. We currently use a dual quad-core Xeon Windows Server box and I wish I had more cores -- when you get a project to OCR 2.1 million docs in a timeframe of less than a few years, you will too. ;-)

        ABBYY's own server-level product (Recognition Server) will span multiple boxes and use any designated cores available on those boxes -- and it scales linerally with the number of cores available (distributed or local). So yeah, there are still some Windows-only applications where a truly monster box would be great.

        OCR is one of those apps where you can absolutely NEVER have enough resources for big jobs.

    • by Macrat (638047)

      Gads, who on earth would run a 64-core Windows box?

      Someone trying to run Microsoft's bloated software?

    • Re: (Score:2, Redundant)

      by shut_up_man (450725)

      64 cores should be enough for anybody.

  • by javilon (99157) on Tuesday September 16, 2008 @09:01AM (#25024023) Homepage

    ... and each one will have it's own processor core.

    • That's actually one of the reasons why Google went with processes for tabs instead of internal threads or trying to use OS-level threads. It's easy to get separate processes onto separate cores.

    • by TeknoHog (164938)
      Yes, because multiprocessor systems didn't exist until Intel innovated the Core. Which raises the question, how exactly did the earlier processor generations do any work if they didn't have any Cores.
  • Being stuck in a windows world I see no benefit in multi core CPUs since XP does NOTHING with the 2nd core. I have a dual core CPU at work and CPU utilization NEVER goes over 50% meaning only one core is doing the work.

    Does VMware workstation or Ubuntu actually make use of the extra cores?

    • by kimvette (919543)

      Then you're running single-threaded applications. Try running multithreaded applications and you can pin both cores at 100%.

    • Re: (Score:3, Insightful)

      by squizzar (1031726)

      In this case, much as it pains me, it has nothing to do with windows, and everything to do with you getting the wrong tool for the job. You have a double garage and two cars, but you can't drive them both.

      I on the other hand have friends I like to let use my spare car: Things like VHDL Simulators and FPGA synthesis tools, that will gladly consume a core and several gigs of memory for a few hours. On a single core machine you might as well go to sleep because the system will be next to unresponsive while

    • Re: (Score:3, Insightful)

      This is a server CPU. Not sure what XP has to do with it, really? People who run servers definitely use multiple cores.

  • by cbreaker (561297) on Tuesday September 16, 2008 @09:53AM (#25024617) Journal

    Until Intel unveils their version of HyperTransport, this will be more of the same.

    You put a quad-core Xeon against a quad-core Opteron and under most conditions (besides CPU-only work) the Opteron will kill the Xeon.

    Now, we'll have even more cycles we can't utilize, because of the old design of the system.

    If you're going to do anything that uses both RAM and CPU (aka VMware hosts, which is what most big servers are used for these days) you'd better off with an Opteron.

    I'd rather use a dual or quad socket Dual-Core Opteron than a dual or quad socket Quad-core Xeon.

    • by mr_mischief (456295) on Tuesday September 16, 2008 @10:21AM (#25025035) Journal

      In the article it's pretty clear this is a legacy-at-launch part to be a last upgrade for people using the current FSB technology. Other products will use QuickPath, but this is for those who want to keep their current motherboards for one more generation of processors.

    • If you're going to do anything that uses both RAM and CPU (aka VMware hosts, which is what most big servers are used for these days) you'd better off with an Opteron.

      You could not be more wrong. Xeons dominate the VMmark benchmarks [vmware.com]. Go home, fanboi.

      HyperTransport may indeed be better design than a shared high-speed bus, but that design advantage is negated by Intel's process and manufacturing excellence. Almost all benchmarks bear this out, SPEC, TPC, VMmark, whatever. Except perhaps at the extreme high end

      • by cbreaker (561297)

        "but that design advantage is negated by Intel's process and manufacturing excellence"

        Who's the Fanboy?

        Look - almost everyone I work with for VMware installations insist on Opteron. Benchmarks are one thing, but real-world performance is entirely different.

        There's no doubt that the Intel CPU's are better performers on single-thread operations or operations where there's not a lot of memory IO. But the Opterons crush Intel when there's a lot of memory IO. Why do you think Intel is moving to a HyperTran

  • Imagine a beowulf cluster of these! =)

    But, can it run Windows Vista? Or how about running Duke Nukem Forever on top of Windows Vista? Or is that just wishful thinking?

  • Can anybody explain me why AMD still has tops 1MB cache per core? And this is on Opterons. Normal AM/AM+ CPUs have only 512K cache per core. Intel now has 4MB of cache per core on its workstation CPUs.

    I can not understand why AMD doesn't increase cache size, while it's obvious that cache helps tremendously on memory intensive operations. Smaller caches are not even offset by its integrated on CPU memory controller.

    I'm planning upgrade and though I like AMD price/performance ratio, even for modest wor

Help stamp out Mickey-Mouse computer interfaces -- Menus are for Restaurants!

Working...