Chimpzilla seems to be planning a campaign to get more of its chips under the bonnets of company servers..
In a prelude to the attack, it seems that AMD has gone headhunting taking staff from both Nvidia and Intel.
The outfit recently hired Intel senior principal engineer, Donald Newell, who according to Rick Bergman, senior vice president and general manager of AMD's Products Group will bring a "strong combination of leadership skills, engineering and design expertise and strategic direction." Well, it is probably more useful than just bringing a strong combination of coffee, water, milk and two sugars.
AMD wants to steal back some of the ground it lost to Intel in the server market. A few years ago it caught Intel by surprise by creating products that allowed it to be actually competitive. All the top OEMs including IBM , HP and Dell adopted AMD and flogged it to corporates.
After a year, Chipzilla caught up and took the crown from AMD again and many are waiting to see what Chimpzilla's next move will be.
Judging by the headhunting that AMD has done, it looks like it will want more mileage from its graphics arm ATI both to provide integrated GPU/CPU chips for laptops but also something more sexy for servers.
Xbit labs seems to think that the new Bulldozer chip armed with a TurboCore feature will be a killer app for the server market.
Anyway, creative naming aside, it seems AMD is now in nearly the opposite situation as it was a few years ago. They're hiring instead of sacking, debts are getting better instead of worse, profit is on the horizon rather than more and more loss. Well, if their plans work out, that is...
Turbo for servers doesn't make much sense. Server workloads are mostly threaded, while turbo is mostly there to boost single core performance... Quite a mismatch. And AMD hasn't been excited about turbo for servers either, but since it definitely benefits clients, why not leave it on for servers too? It might just help catch some load peaks if implemented nicely.
I personally have little doubt Bulldozer will do well on the server market. It's designed for throughput per Watt and per mm^2 (read: product cost), which is exactly what that world seems to be all about. The only things that could really screw stuff up are a dodgy process, an architecture bug and for too high expectations (even if they don't originate from AMD) - a familiar sounding combination, no doubt.
A-what? Do you even read what you link to or do you read it but still decide to make make an fraudulent claim? Xbit never commented on the news, but solely wrote up what looks like a press release.
"recently hired Intel senior principal engineer,"
They recently hired an ex-intel engineer. Jesus.