Two AMD news items haven’t gotten much play, but they are important:
It’s hard to say whether the first item is a change of plan or just vocabulary. AMD has previously said that high-K was “an option,” now they “plan” to do it. Mind you, they didn’t say they were going to do it; they just said they planned to do it. You can change/end a plan to convert a lot easier than change/end a conversion.
I suspect they really would rather do it, probably because the “cheap” 45nm process is no barnburner, it’s a question of whether they’ll have the money to pay IBM for it.
AMD has been very, very quiet about Bulldozer lately, so it’s nice to know it hasn’t been shuffled off, but . . . .
Anyhow, the first Bulldozer samples are supposed to show up sometime in 2009, and real ones will show up ???. Remember that AMD initially promised this in 1H 2009 back in July 2007, but dropped that idea a few months later, apparently due to lack of development money.
Given the history, it’s probably safest to assume that any Bulldozer anybody sees will not be a brand-new architecture but instead be closer to a tweaked K10.5, with high-K providing the bulk of any performance improvement, assuming they end up being able to afford high-K, of course. AMD doesn’t have the time nor resources to do more than that.
If all this strikes you as being tentative, too tentative, it is. What we see here are the plans of a financially crippled company, a company laying out plans without the cash to pay for them.
I’m not saying these plans are complete fantasy, not at all. They are the plans of a company that knows its real #1, top-priority mission is to get more money, somehow, someway.
I know a lot of eyes glaze over when I talk about finances; they want to read about computing, not accounting. This is understandable, and so long as a company makes enough money to pay for its plans, it’s quite justified. To the average person reading this, it really makes no difference whether Intel made $1.5 or $2 billion last quarter.
It’s only when finances adversely affect a company’s ability to compete today and develop tomorrow’s competitive products that it matters to the average person reading this, and that’s exactly the case with AMD. At AMD, money (more precisely, lack thereof) is the 800-pound gorilla in the room. They don’t like to talk publicly about it, but it’s not a factor in AMD decision-making; it’s the factor.
Lack of money has always been AMD’s Achilles Heel, and now it’s worse than ever.