A few websites got to tour the place. Wonder who paid for the trip. 🙂
The timing of this trip should raise a few eyebrows. AMD will release its financial earnings tonight. The overall numbers for the quarter will be thw worst ever simply because AMD will be taking a big charge for its restructuring. The important numbers will be those showing AMD’s operational loss, which should be much better than last quarter. We’ll talk about those tomorrow.
However, most financial bozos will just report that single number, which will generate bad press. This arranged trip is in all likelihood a PR attempt to get some good news circulating.
Most of these articles describe the facilities, but I suspect most of the audience is far more concerned with what is being made and how well it does rather than how they’re being made.
The one area of general interest that apparently was broached by those on tour was Hammer.
Of course, the comments made by AMD about Hammer are just plain silly, “It was doing so well, we redesigned it! That was so much fun, we delayed it another six months!”
Spare me. If you believe that, I have a bridge to sell you.
All the talk in the world about yields means very little if you don’t know what the speed bins are. You can have 99% Hammer yields, but if they’re all 800MHz processors, it doesn’t mean much, does it? That probably was the situation the first half of the year.
What would be a far more likely description is, “We couldn’t get them to work fast for most of the year, so we had to go back to the drawing board ala TBredB.”
It looks like AMD has improved since then, but how much more so? We’ll find out next summer.
2GHz Is Too Late For Overclockers
The articles pointed out that a 2GHz Clawhammer will be the rough equivalent of a PIV 3.06GHz with hyperthreading, which is roughly about what we have expected all along.
The problem for AMD in the overclockers’ market isn’t even Intel. It’s that the latest TBreds aren’t too far away from that level today, they cost just $100 today, and certainly will cost less six months from now, when AMD will start seriously cranking Clawhammers out.
It’s going to be hard to persuade people to leave their socket A boards in 2003. Right now, there seems to be a substantial group of people out there who think $100 is too much for a CPU; $55 to $60 is more like it.
Even those not quite as thrifty may blanch when the choice is between a $70 processor upgrade and, say, a $350-400 CPU/motherboard upgrade six months from now. Especially with .09 micron chips and DDR-II memory on the horizon.
It seems to me you’re going to need to be able to get about 2.5GHz out of this generation of Clawhammers before it becomes a really enticing proposition to overclocking socket A owners. And it had better be pretty cheap.
It’s not inconceivable, and there are indications this will be doable at some point in Clawhammer production, but when?
The later it is, the better the argument for waiting until 2004 for .09 micron chips.