Intel preps 45nm Xeons for November

Word on the street is that September 10 will be the day AMD rolls out its first 65nm quad-core “Barcelona” Opterons. These processors will naturally have to prove themselves against Intel’s 65nm quad-core Xeons, but if a new report by DailyTech is correct, they’ll also be facing competition from 45nm Xeons within just over two months of their release.

DailyTech says Intel is cooking up seven quad-core 45nm “Harpertown” Xeons for a launch on November 11. The chips will have 2GHz, 2.33GHz, 2.5GHz, 2.66GHz, 2.83GHz, 3GHz, and 3.16GHz clock speeds, and they’ll all pack 12MB of L2 cache. All chips except the 3.16GHz model will have thermal ratings of 80W, too. Prices will range from $209 for the 2GHz model to $1,172 for the 3.16GHz one. For reference, the Barcelona pricing information that leaked yesterday suggests the 2GHz quad-core Opteron will cost $390 when it launches next month.

Comments closed
    • blastdoor
    • 13 years ago

    too bad Xeons are hobbled by those expensive, power hungry FB-DIMMs. If it wasn’t for that issue, AMD wouldn’t have a leg to stand on.

      • shank15217
      • 13 years ago

      however Intel was obviously aware of the power consumption of fb dimms and decided to use it anyways. AMD made a smarter choice as power density in server rooms is one of the highest criteria. Intel has to live with their choice for a long time.

        • Krogoth
        • 13 years ago

        It is because FB-DIMMs are the only way Xeon can compete against Opterons in memory capacity.

        It will once CSI and intergrated memory controllers get onto future generation of Xeon processors. 😉

          • tfp
          • 13 years ago

          Ah yes and that should be around 2008. It will be interesting to see what they do cache wise once they have an on board memory controler.

    • Vaughn
    • 13 years ago

    I welcome the upcoming price war!

    • Dposcorp
    • 13 years ago

    Does anybody at Intel or AMD work harder then the Press Agents?

    • Anomymous Gerbil
    • 13 years ago

    What does it mean when they say “All chips except the 3.16GHz model will have thermal ratings of 80W”… do they just pick the highest ceiling, even thoguh the lower speeds chips will presumably disspiate a lot less than 80W? Why not specify a different level for each chip?

      • cobalt
      • 13 years ago

      My assumption is that the vendors have to certify their systems for a given TDP, and more levels means more work for them.

        • Anomymous Gerbil
        • 13 years ago

        Maybe… althoguh they could just certify for the highest setting and be done with it? Then the various TDP figures would tell users what the worst-case power usage is likely to be.

          • jdevers
          • 13 years ago

          Yea, but there are probably lots of OEMs that won’t go above 80W on some of their lines. Also, there is a good chance that TDP is very non-linear and the top end chip outputs considerably more than 80W.

            • Anomymous Gerbil
            • 13 years ago

            Probably true.. but I’m talking about the chips that are *lower* speeds in the 80W range, and therefore probably lower TDP. Just curious as to why Intel and AMD group a bunch of chips under the highest common TDP, instead of specifying them individually.

            • dragmor
            • 13 years ago

            Because TDP is not meant to be a measure of each individual chips power consumption. TDP (Thermal Design Power) is a measurement provided to the OEM to say design a PC that can cool a 80w CPU and all of these CPUs will work no testing required.

    • droopy1592
    • 13 years ago

    12?

      • cobalt
      • 13 years ago

      What, the cache? Sure, why not — the 65nm ones already have 8MB, so that’s not a staggering increase. Plus, with the memory controller off-chip, it’s still definitely useful.

Pin It on Pinterest

Share This