WOW!!! . . .

First, from DailyTech:

“One vendor demonstrated Barcelona to DailyTech running at 1.6GHz. Current AMD Barcelona samples are not scaling too well. AMD partners confirmed the highest running, POST and OS capable, Barcelona processor is 2.0 GHz. AMD previously posted benchmarks of a simulated 2.6 GHz Barcelona.”

What can you say? This is not a competitive chip at 1.6GHz. Now the Povray scores make sense (it should be noted that current AMD chips are pretty lousy at Povray, anyway). Some intrepid soul did manage to run a benchmark on the Barcelona machine before they “were instructed to leave,” and well, they earned the page hit.

We’ll just note that:

  • you’ll see the actual Cinebench scores in the screen shots,
  • the bench used was Cinebench 9.5, so no SSE4 finagling here in some beta release,
  • Cinebench pretty much scales according to processor speed, so
  • clock-for-clock, the Xeon tested is about 5% faster
  • it’s certain Intel’s Penryn chips will have rather higher clockspeeds than Barcelonas, both sooner and later.

    If you take another look at current processor scores and do a little estimating, it looks like the Barcelona improvements only narrow the gap against current Intel processors by a few percentage points.

    More importantly, we do know what Penryns will do at 3.33GHz to this benchmark.

    Assuming perfect scaling (and Cinebench pretty much does that), a Barcelona would have to run at 3.8GHz to match a 3.33GHz Penryn. I repeat, not 2.8, 3.8GHz. Given that the top speed of the initial Barcelonas looks to be (hope, hope, pray, pray) 2.3GHz, a quad-core Barc at that speed would only be about even with a dual-core Penryn.

    If Cinebench is representative of what the Barcelona generation can do (and all the other indicators we have are pointing in the same direction), what we have is a slightly more tweaked Hammer, one that doesn’t even look as tweaked as Penryn. Even if they get the clock speeds up to the levels promised, we’re looking at catchup with C2Ds, then Penryns will reestablish the performance lead.

    We’ll need more benchmmarking, of course, but even if things go well, it looks like Intel will retain the current pricing stranglehold (maybe with a small interruption). AMD just cut prices on Opterons from 10 to 50%. It will cut desktop prices another 15-20% next month.

    Frankly, for AMD, it looks like a matter of what happens first, do they somehow make some money, or do they run out of it?


  • Be the first to comment

    Leave a Reply