Abit fiddles while the PC-enthusiast world moves on............

Discussion in 'Abit' started by John Lewis, May 13, 2005.

  1. John Lewis

    John Lewis Guest

    I notice that MSI's P4N Diamond Socket 775(T) Intel motherboard is
    already IN STOCK at NewEgg @ $239.

    What's "wonderful" about that ?

    * nForce4-SLI INTEL chip-set, support for both single and dual-core
    Intel processors, with dual PCIe graphics slots. On-board audio is
    24-bit Creative Soundblaster 7.1-channel on the version Newegg
    currently stocks ( beware, an apparent alternate for this board is
    Realtek.... yukkk... )

    * SOFTWARE switching of SLI/non-SLI modes. The horrible kludge-card
    is gone.

    So you can use your current LGA775 Processor in this board, configure
    your PC for 4-head dual-card ( for graphics-arts/video ) or
    2-head-SLI (for gaming) whenever you wish and without opening the
    case, and upgrade to dual-core at your convenience.

    Not that I recommend the Intel dual-core-CPU solution at all, since it

    consumes 60-90 watts more than the corresponding-performance
    AMD X2 solution --- which unfortunately may only be available in
    retail channels late Q3 2005.

    The current motherboard offering on the AMD-939 single/dual-core
    side that supports a SLI software-switch is the A8N-SLI Premium, but
    this does not seem to have reached any retail channels yet.

    Seems as if Abit is more than just a little behind the times in the
    enthusiast market on both the Intel and AMD side, since there
    has been zero announcement of any Abit motherboard with software
    configuration of SLI, let alone shipment. Too busy fiddling with
    gimmicks like uGuru and OTES no doubt............

    John Lewis
     
    John Lewis, May 13, 2005
    #1
    1. Advertisements

  2. "John Lewis"...

    Yawn John. Isn't it time you changed the record?
    --


    Richard Hopkins
    Cardiff, Wales, United Kingdom
    (replace nospam with pipex in reply address)

    The UK's leading technology reseller www.dabs.com
     
    Richard Hopkins, May 13, 2005
    #2
    1. Advertisements

  3. John Lewis

    John Lewis Guest

    Most unlike you not to give us some deep technical insight into the
    background for your yawn.........

    On a related topic....

    Regardless of Abit's (and your unexpected) lethargy, Intel has a
    teeny-weeny (?) architectural problem with dual-core, specifically
    related to its excessive power-dissipation. Intel's marketing machine
    will be working overtime to smooth the cracks and Intel's huge
    production machine will turn out dual-core at fire-sale prices to keep
    AMD out until they solve their architectural problems.... maybe
    synchronous with the move to 65nm... although some little birdies I
    know at Intel say that particular process-move might be at least as
    difficult as the move from 130nm to 90nm. Seems as if Intel should
    actively consider shipping a CPU water-cooling kit instead of a
    fan/heat-sink with their current-generation boxed dual-core
    processors.....at least for desktop applications.

    I happen to live not far from Intel's advanced architectures group and
    have some close personal friends working there. They do not divulge
    any specifics of their work (and I do not ask them either) , but they
    have expressed disquiet over the past few years about a management
    style that is excessively cautious about innovation in CPU
    architectures. And a touch of not-invented-here about anything
    innovative that AMD comes up with. This may have been the path that
    led to the Prescott fiasco and the current kludge
    dual-core-Northbridge CPU architecture that Intel is now hurridly
    trying to fix.

    John Lewis
     
    John Lewis, May 13, 2005
    #3
  4. Well, if you insist. It's just another of your regular "Abit aren't making
    the products I'd like to see them make" whines. As always, the solution's
    the same - if you like another manufacturer's sh*t better, go buy it.
    No, the power dissipation problem is nothing to do with "dual core" itself.
    The power dissipation (or more accurately, current leakage) issue is a
    Prescott goof, although it certainly does knock on to the Smithfield, which
    is of course nothing more trick that two Prescotts in a single package
    Your point being?
    Yes and no. The 65nm Netburst chips don't suffer from Prescott/Smithfield's
    current leakage problems, and do have a fair bit of clock headroom. However
    the real "fix" to Intel's current mire comes with the Merom family either
    late this year or early next, depending on when they decide to put them out.
    No. Intel's 65nm process is working very well by all accounts, none of the
    problems that bedeviled the previous transition. Seems your birdies aren't
    tweeting the right tunes, or maybe they're a little out of the loop.
    They're not that bad. The newer stepping seems to be a little less
    extravagant under load, and the TM2 helps off-load. Even so, the 65nm
    Presler and Cedar Mill chips are much better, pulling around half the power
    of an equivalently clocked Prescott or Smithfield
    Most of the important R&D work over the last couple of years seems to have
    been done at the Jerusalem facility. That's where Merom and its siblings are
    coming from
    That's fair comment, and anyone can see what a f**k-up Netburst was. A
    triumph of marketing over engineering. However, its failure has, if nothing
    else, enabled the engineers to get the upper hand right back. The Merom
    family kicks ass, both on its own terms, and in terms of competition to AMD.
    The "Prescott fiasco" as you accurately call it was a result of a number of
    different things, from bad product planning to poor project management to
    over-enthusiastic design, to production problems.
    Hurriedly? The solution has been in development for over three years. It's
    precisely because they haven't hurried it that we've got Smithfield, which
    is, even with the best will in the world, a kludge. Still, the 65 nano line
    will reduce the rot, and what comes next will dramatically alter the balance
    of power.
    --


    Richard Hopkins
    Cardiff, Wales, United Kingdom
    (replace nospam with pipex in reply address)

    The UK's leading technology reseller www.dabs.com
     
    Richard Hopkins, May 14, 2005
    #4
  5. John Lewis

    John Lewis Guest

    without an on-chip memory controller and cross-bar, all of which save
    power by not having to waggle lines on the motherboard.
    Regardless of your later comments, Intel is under severe internal
    pressure to move past the current generation of dual-core asap,
    specifically to get the power-dissipation under control. AMD
    will capitalize on this Intel Achilles' heel in their X2 and Opteron
    marketing. The Catch 22 question for Intel is: do they go into full
    production on 65nm with a new multicore CPU architecture and possibly
    fall flat on their faces again with unanticipated process problems, or
    put just the dual-core version of their new architecture asap on
    90nm.... regardless of their fancy marketing projections and plans.
    AMD is looking to a smooth transition to 65nm multicore when
    the time is right and in the meantime on 90nm has a very
    power-efficient Opteron multi-chip solution for those that
    need more than 2 cores.
    My birdies are quite knowledgeable.. the process itself is only part
    of the issue in making the transition to the smaller geometry and then
    cranking up volume production......

    For any process with only a few hundred electrons on each gate, it is
    a devil trying to establish statistically-accurate transistor and
    complex-logic-timing models, with noise and other stray-effects
    becoming an increasingly-important factor. Thus, the logic models
    become increasingly more like analog-circuit models as the processes
    shrink, hence the computation-load grows exponentially. Intel is
    currently massively updating their design server-facities to handle
    the simulation work-load of their next-generation processors. The
    implementation of massive architectures on brand-new and
    smaller processes takes significant effort and time, with many
    hiccup possibilities on the way...........
    will be.........
    Design and Development maybe, not architecture research........

    John Lewis
     
    John Lewis, May 14, 2005
    #5
  6. John Lewis

    No One Guest

    I have a Guru card. It sucks.

    As for SLI, nVidia only show it working with 1 monitor. What gives?
     
    No One, May 18, 2005
    #6
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.