Let me respond to you why your advices would've resulted in bigger crisis than AMD is in now.
BaronMatrix :
You mean a quad core 45nm laptop is a bad thing? I wish I was running AMD sometimes. I wouldn't have dropped prices as much and would definitely have put more resources into mobile. They should take my advice and shrink Griffin with Stars cores and forget about being able to sell $1500 desktop processors that no one buys anyway (what about .2%).
Well at the moment, Intel is the only one with a quad core mobile CPU, and that unit itself costs over 1000 USD. Aside from that, AMD's CPU would have to lower mobile Deneb's prices in order to compete aganst the mobile Yorkfield. How many people will actually spend that much money in purchasing such a high end mobile CPU? As statistics show, people are much more willing to pay for high end graphics than CPU. If that's the case, what's the point of AMD entering into the fray, commit resources in optimizing Deneb for mobile, and further divert resources from else where? It also seems that Deneb will not have the TDP to actually fit into the mobile segment. AMD would have to wait for 45nm HK/MG, or even 32nm HK/MG before attempting something like that.
I'm not sure how far you would've dropped the prices, but if at the moment performance / price is the only reason why people would select AMD over Intel. In the world where switching from one company to another means an entire makeover in your computer components, AMD's first goal is to secure a relatively safe footing in the market share; to keep AMD users from switching over to Intel. Even if they do lose money now due to extremely low prices, the platform compatibility will allow users to purchase much higher performance CPUs, with much better margin, in the case they do come out.
As for mobile, you have to realize one thing: mobile sector is Intel's playground. Intel has been dominant in the mobile sector since the launch of Centrino back in 2003. Centrino laptops have became a household name for laptop. Therefore in order for AMD to penetrate the market and establish a footing, they would not only have to commit extra resources in altering the current AMD system architecture to favor mobile platform, but also spend enough on marketing to increase the brand awareness. For a cash strapped AMD, it makes much more sense to be in a defensive position, and fortify its foothold in the server segment.
If I wereto give Intel advice it would be to have left ATis chipset license and stop tying products together. What happens is that if they have a problem, where will people get chipsets or IGPs for all those CPUs? I would also have stayed out of the XO market. Intel is stretched too thin, almost like MS. Nehalem may be their Rubicon as it will require new boards, new chipsets, a whole new infrastructure. I'd also wait and drop server chips first, since if the plan backfires i7 will sit while Deneb and Penryn sell out.
I'm not exactly sure about what you meant by "typing products together", but your "suggestion" of Intel leaving ATI license couldn't be more wrong. Due to the lackluster performance current AMD CPUs have, and users tendency to go for Intel's offering, by confining Crossfire technology only to AMD chipset is no doubt a suicide, just like Nvidia. AMD would lost a lot of graphics card sales just because Crossfire technology is not available on Intel chipset, which is currently the best chipset available on the market for Intel CPUs.
As for IGPs, I think you're overestimating the importance of graphics power. As statistics show, Intel currently controls about 60% of the IGP market, while AMD only controls ~15%. However everyone + dog knows that Intel's chipset sucks arse, while AMD's chipset kicks arse. Why is that?
Personally I don't understand why you have a beef with Intel entering the XO market. Intel is not stretched too thin.
Don't worry about Nehalem. It will do much better than you expect. Sure, it requires a new board, new RAM, but it also offers a lot of advantages that Penryn and Deneb cannot offer. It won't be too long before OEMs switch to GFX equipped Havendale for mainstream and low end PCs, to conserve cost by taking off redundant chipsets.
On the other hand, it would make absolutely no sense by releasing server parts first. Remember Dunnington? It has just been released this year. By releasing server Nehalem parts will only eats up the sales of those products. Oppositely, if Nehalem is released to the enthusiasts crowd first, then it may build up the hypes, which in turn helps more companies to switch to Nehalem.
Facebook
Twitter
Instagram