|
|
|
|
|
|
|
|
|
||||
|
|
|
|
|
||
|
|
|||||
|
|
|
"AMD Thoughts"
As you say AMD, is in a world of woe at the moment. As an AMD user I can suggest why. Any sales of the current AM2 lineup has got to come from one of four groups:
But there is one way that they could survive the next couple of months using technology they have now that would be very attractive to end users at a fair price - not fire sale, nor silly bleeding edge pricing.
What do they need to do to save the company?
Take the DDR memory controller and graft it back onto the 65nm CPU's and use it to pump out some Socket 939 5000+ and 6000+ CPU's at a reasonable price. They already have the DDR controller technology and they already know how to connect it to the current architecture - which hasn't really changed from 939 anyway.
Why do it? Those millions of Socket 939 loyal fans would love them (buys loyalty), helps to keep a customer base. If I had the choice between a new board + new CPU + RAM or a new CPU that brings me back up with current speeds, I'd buy the new CPU and maybe put the money I was going to spend on the rest to a new video card that AMD will want me to buy when they ship the X2K series.
The 939 boards out there have aging CPUs in them that are all pretty much >12 months old, so the users are edgy; but faced with going to AM2 and getting burnt again or waiting for K10, we'll wait - or go Intel. A new CPU in the current board - especially if it looked like a limited proposition - would move some CPU's for a couple of months to keep them in the game.
Now extend the Fusion/Torrenza concepts and add a selectable memory controller so that a K10 / Quad Core 939 chip is a reasonably easy to make option, and AMD could bounce back.... But will they?
|