Leaked Intel Nehalem performance projections over AMD Shanghai

Leaked Intel Nehalem performance projections over AMD Shanghai

Summary: It appears that the rumors about Intel's next major microprocessor "Nehalem" being a huge juggernaut may be true according to leaked documents from Sun Microsystems (removed Sunday night).  The slides appear to be inadvertently placed on Sun's publicly accessible website and "jokerman" posted the link on Aceshardware (thanks to tip from ZDNet reader JumpingJack).

SHARE:

It appears that the rumors about Intel's next major microprocessor "Nehalem" being a huge juggernaut may be true according to leaked documents from Sun Microsystems (removed Sunday night).  The slides appear to be inadvertently placed on Sun's publicly accessible website and "jokerman" posted the link on Aceshardware (thanks to tip from ZDNet reader JumpingJack).  The slides looks like the real thing meant for Intel's partners and they're probably well known in the server industry.

Reliable sources have reported in the past that Intel's Nehalem processor will have three channels of DDR3 memory per CPU versus two channels of DDR2 memory per AMD Barcelona or upcoming Shanghai processor.  That would mean that AMD's massive memory bandwidth advantage will turn in to a large memory bandwidth.  So what does this mean for Intel Nehalem's performance?  Take a look at the following charts I generated after carefully measuring the length of the performance bars on a pixel level.

Since Intel's charts were normalized to an Intel E5160 dual-core processor on SPECint_rate_base2006 and SPECfp_rate_base2006, I had to start somewhere and make some guesses on the base performance.  I used Intel's highest published SPEC CPU integer and floating point score of 60.8 and 45.1 for the E5160 processor as of 2/23/2007.  This is probably not the exact reference point that Intel used so the numbers might be off a little.

When I compared my extrapolated numbers to the published SPECint scores for all of the shipping products other than the E5160, I found that Integer performance was 2% to 7% too low and the average was 4%.  When I compared with published SPECfp scores other than the E5160, I found that my extrapolated numbers were all 4% too high for all models except the Opteron 2220 extrapolation which was 12% too high.  To adjust for this, I raised the SPECint estimates 4% and dropped the SPECfp estimates 5% and generated the following chart which is a closer match to the published scores.

I tend to believe that the second adjusted chart is more accurate.  We'll most likely know by the end of this year what the actual scores are, but I doubt they will be more than 5% to 10% off from these estimated projections.

So how can Intel pull off such a massive performance boost over their current reigning champion "Harpertown" X5482 processor?  Consider the fact that Intel's current generation 45nm Harpertown processors lead the benchmarks despite the memory bandwidth disadvantage because of a much faster execution engine and larger cache.  Then we factor in the fact that Intel will implement SMT (dual threads per CPU core), improve the already-fast execution engine of Harpertown, and feed it with three channels of DDR3 memory per CPU instead of the old shared front side bus.  AMD's Shanghai on the other hand is essentially a die shrink, a cache size boost, and a clock speed boost.  Taking all these things in to consideration would easily explain how Intel could widen the lead so far.

I would also note that Intel's leaked slides compare these processors in pairs where the Opteron 2220 DC (Dual Core) faces off with the Intel E5160 DC processor and the 2222 faces the X5365.  These two pairs represent a snapshot in time to when the products competed against each other.  The last two pairings on top may be generous to AMD since Barcelona processors aren't shipping yet because of the TLB bug whereas Intel launched the X5482 in November 2007.  AMD's Shanghai processor didn't have first silicon until four months after Intel showed off their first silicon at spring IDF 2007 in September, but the difference is that Intel has showed the Nehalem running a real Operating System while AMD has not done the same for Shanghai.

Since it usually takes one year from first silicon to production parts, it's a bit hard to believe that Shanghai will ship at the same time as Nehalem.  But even if it does ship at the same time as Nehalem, the competition from Intel looks very daunting if these estimates are anywhere close to being accurate.

Topics: Hardware, CXO, Intel, Processors, IT Employment

Kick off your day with ZDNet's daily email newsletter. It's the freshest tech news and opinion, served hot. Get it.

Talkback

95 comments
Log in or register to join the discussion
  • Lies, lies and lies

    """
    Barcelona processors arent shipping yet because of the TLB bug
    whereas Intel launched the X5482 in November 2007.
    """

    You know just as well as me that as of today it is
    [b]impossible[/b] to buy the X5482 processor (except if you buy
    a Mac Pro because Apple has priviledged relations with Intel).

    Whereas, despite the TLB bug, you can find Barcelona processors
    at pretty much every big reseller: newegg, amazon, etc.

    How can you knowingly misrepresent the reality like that ?!
    zpdixon 42
    • Dell offers the X5482

      zpdixon said: [i]"You know just as well as me that as of today it is
      impossible to buy the X5482 processor (except if you buy
      a Mac Pro because Apple has priviledged relations with Intel)."[/i]

      Not so.

      http://configure.us.dell.com/dellstore/config.aspx?c=us&cs=04&kc=6W300&l=en&oc=bwdwjap&s=bsd

      zpdixon said: [i]"Whereas, despite the TLB bug, you can find Barcelona processors
      at pretty much every big reseller: newegg, amazon, etc."

      Not so. The following are search results from the sites you mention:

      http://www.newegg.com/Product/ProductList.aspx?Submit=ENE&DEPA=0&Description=Barcelona
      http://www.amazon.com/s/ref=nb_ss_gw/002-1949112-9469664?url=search-alias%3Daps&field-keywords=amd+barcelona
      John Mathers
      • Uninformed FUDsters...

        ...are AMD fanboys.
        thetruthhurts
      • Uh oh, looks like you need to learn how to use the 'search' feature:

        http://www.newegg.com/Product/Product.aspx?Item=N82E16819103225&ATT=19-103-225&CMP=OTC-Froogle&nm_mc=OTC-Froogle&cm_mmc=OTC-Froogle-_-Processors-_-AMD-_-19103225
        http://www.amazon.com/gp/product/B000YPW6IU
        http://www.buy.com/prod/amd-phenom-9600-quad-core-2-3ghz-4mb-cache-95w-tdp/q/loc/101/206359022.html

        All in stock.
        zpdixon 42
        • You forgot

          Agena != Barcelona ergo the point still holds. You can't buy Barcy.
          SuperFly03
        • Hey, didn't you misrepresent things in the first post?

          AMD's PR department must be really lacking these days
          thetruthhurts
        • Ohhh, you don't know what subject you are talking about.

          Ohhhhh... you seem to equate markets of server to DT.... these are Agena not Barcelona ...

          Server cannot tolerate even the most minor glitches, downtime and lost data costs money... AMD doesn't care if you loose your latest changes to your last love letter to susie rottencrotch when it locks up, you would probably blame Vista bugginess anyway. They are more than willing to send out what the enterprise does not want onto market for the average joe to use up.....

          Now, having said that.. the chip is stable enough... AMD did the right thing, the frequency of lockup is not high enough to be of concern to the average joe.
          JumpingJack
    • true, true, true

      X5482 can be had and found, sever systems and parts are not as accessible in advertisement as your common desktop or laptop. Simply call around and ask....

      On your second point, Barcelona had a brief showing at those etailers you mentioned late september, but then were removed from the shelves. Though you can find some OEM rejects at a few no-name places, they are essentially non-existent. AMD is only shipping barcelona on a case by case basis, after they can ensure that the TLB bug is irrelevant. That news was big headlines last december, for example. http://arstechnica.com/news.ars/post/20071212-following-up-on-phenom-performance-and-availability.html

      Of course, to save any humiliation, AMD said they really did not do a stop shipment.http://www.theregister.co.uk/2007/12/06/opteron_delays_amd/ more or less saying 'you cannot stop shipment on something that has not shipped.' Oddly, they launched the friggin' thing Sept 12, 2007 with great fan fair at Lucasfilm with oodles of people heaping praise on a processor that, technically, won't ship until March or perhaps even April of 2008....

      The Barcelona debacle has been amusing to watch, and sad how badly it has been botched.
      JumpingJack
    • Another lie

      Here are a couple places where you can get the Xeons X5482

      http://www.rackmountpro.com/product.aspx?prodid=2089&catid=217
      http://www.memorylabs.net/inx5bueu.html

      That's the reality.

      I didn't bother checking other places.

      Who misrepresented "the reality" again?
      thetruthhurts
  • AMD is the top performer

    Sorry dude, but the reality is that AMD chips perform a lot better than Intel chips at the same Hz.

    If you compare two CPUs at different speeds w/ different cash memory sizes, you are not comparing performance on the equal level.

    But when you compare two chips with the same exact (real) speed and same cash size, you will find that since the Athlon came out, Intel chips have always being inferior in performance and power consumption. Most AMD CPUs either outperform or match the performance of Intel CPUs with 1 GHz more. Actually a 2GHz Athlon 64 X2 matches the performance of an Intel Core Duo running at 5GHz.

    I understand that for the simple minded person, Intel looks like a better performer. After all they do have the fastest chips in the market (not necessarily the best). But REAL performance should not be measured by who is the fastest, but how is doing more for less. In that arena AMD is winning by a huge leaps.

    And BTW, for those who think the AMD Phenom (Quad-Core) is buggy, check out the 5 page table of "NO FIX" errata (w/ 31 pages of explanation) for the Intel Core Duo:
    ftp://download.intel.com/design/mobile/SPECUPDT/30922214.pdf
    wackoae
    • AMD lost the crown in June 2006.,

      >> Sorry dude, but the reality is that AMD
      >> chips perform a lot better than Intel
      >> chips at the same Hz.

      Clock for clock Intel has been ahead since the Core 2 came out in the middle of 2006. Even the release of the Barcelona/Phenom hasn't changed that. "Sorry dude" you need to go back to the times of AMD's K8 vs Intel's Netburst for a time when AMD's chips performed better at the same clock speed.
      bruce2003
      • Almost, not quite

        AMD still has a lead on SMP applications, because of HyperTransport and their dedicated memory controller. Unless a workload fits entirely in cache, the shared bus on the Harpertowns runs out of steam somewhere between 4 and 8 cores. On most benchmarks, the 8-core Mac Pro does no better than the quad-core:
        http://www.barefeats.com/octopro1.html

        The introduction of on-board memory controllers and QuickPath Interconnect (QPI) in Nehalem will remove the last hindrances to Intel regaining the crown across the board. That's why I am holding out on upgrading my Dual-G5 PowerMac to a Mac Pro until the Nehalem Mac Pros are out at the end of the year or beginning of next year.
        fazalmajid
        • What...are...you...smoking?

          Go look up TPC, SAP, SPEC, etc scores.

          Come back when you have and tell us again.

          Do you know what SMP stands for?
          thetruthhurts
        • Yes/no in some cases

          Your statement is generally true only in cases of large working sets that are needed for multiple memory calls (which are usually HPC related, and some transactional databases). The other reply is correct.... the general run on the data is that Intel is ahead substantially, with a handful exceptions where the BW argument is the hinderence.

          However, for the volume server class applications, Intel has pulled ahead and the results show. AMD has been able to hold (even take) some market share in the consumer segment (they are practically giving their processors away), but since the core architecture debuted they have lost almost 1/2 of their market share back to intel going from 25% range to the 12-13% range. IT procurement know what is what.... the change in marketshare typically shows the correct pattern... just as when opteron vs netburst was in play, market share went AMD's way.
          JumpingJack
    • Message has been deleted.

      SuperFly03
      • It takes a bigger moron ...

        ... to call somebody a moron.

        Just check the fact. A 2GHz AMD CPU will outperform a 2GHz Intel CPU any time of the day. That is a fact.

        If you are too stupid to understand that simple fact, then please go back to school and get an education.
        wackoae
        • Please check your facts before speaking

          That is in no way true. During the Netburst era AMD did have an IPC advantage but that was erased with the introduction of the Core microarch, I don't care what time of the day it is.

          I'm not the ignorant one around here. You are the one who is misinformed and hard headed and a complete AMD Fanboi.

          http://www.tomshardware.com/2006/07/14/core2_duo_knocks_out_athlon_64/page15.html

          A slower Core 2 beats AMD, so put that in your pipe and smoke it. Bring facts to the table next time.

          If we go the server route that is an entirely different set of circumstances.
          SuperFly03
          • Superfly003

            Wow, flys really are unpleasant little insects....what an appropriate name.
            nickoli0_z
          • Insults?

            What's wrong? You can't handle the truth so you resort to petty insults? Every review site will show that Intel has a commanding lead in all market segments: desktop, mobile, SP servers, DP servers and MP servers.
            GIANTOYSTER
          • Next time

            At least get my name right. Also, check the facts. If you get pissed about the truth then you don't belong here.
            SuperFly03