Someone sounds a bit overzealous for the boys in blue
Hah, you wish
Actually, I've never really liked Intel at all, and I think x86 is the worst thing that ever happened to the computer industry (therefore I hate AMD for introducing their 64-bit extensions, and dragging on the x86 legacy for at least another 10 years).
No, my favourite CPUs are the 68000, PowerPC, PA-RISC and Alpha.
But I have to be realistic, it's an x86 world (and a Windows world for that matter)... So I too have to use x86 processors for normal desktop/notebook systems. And as an objective 'outsider' to the x86, I can draw no other conclusion than that Intel is currently doing better than AMD, and also has better plans for the future than AMD.
I can't speak for what the performance will be for Nehalem, but Bulldozer will be significantly faster than Barcelona, both in single thread and multi thread situations.
Uhh yea, it had better be. It doesn't take a genius to figure out that Nehalem will also significantly improve on IPC, especially in multithreading, because of its single-die design, integrated memory controller, and better multi-socket infrastructure (basically exactly the things that worked for the Athlon64, so especially you guys at AMD should know how significant this change is).
Considering the fact that Barcelona can barely keep up with Kentsfield, AMD had BETTER get significantly better performance out of their next-gen products. Else you might aswell close shop right now.
45nm is on track, as is Shanghai. Sandtiger follows shortly after. Bulldozer based Fusion processors are due for 2009.
Yea, wow 'on track'... Meaning they're still not gaining on Intel in terms of production process.
And what market exactly is this Fusion supposed to address?
Who's going to want it, and why?
It sounds like a niche-product, not the cash-cow that AMD needs.
And do you honestly think Shanghai or Sandtiger will be cash-cows? It all seems like the same kind of bland rehashes of technology as Barcelona is. Now Penryn is also a rehash, but that's a rehash to make a very successful architecture even cheaper to produce, and will only be used for a year, to hold Intel over for Nehalem.
AMD needs more than a rehash, they need something to turn their outdated product-line into something new, fast, desirable and successful. Where is AMD's 'Nehalem'?
Oh, and may I also remind you that Intel has decided to step up the Itanium-action, and no longer trail a process behind? This means Intel has an ace up its sleeve if things really turn out to be a multicore war. Itanium cores are very tiny, you can fit about 60% more Itanium cores into the same transistor budget currently. When Intel gives Itanium the same transistor budget as their x86-line, we could end up with a whole lot of cores on a chip.
And, what is AMD going to do about the poor Radeon line? That's another product line that needs more than some rehashes to get back on track.
In summary, I think you are selling AMD short
No, I think we are sticking to the exact same facts. I just draw conclusions from them, you seem to just count the number of things you can find on AMD's roadmap.
Now I've explained why I don't think AMD has any kind of technology on that roadmap which indicates that AMD is going to leapfrog Intel.
If I'm wrong, please point out what technology is going to make AMD leader again, and why. In detail please, I'm a hardcore tech, I don't go for sugarcoated marketing nonsense ('native quadcore!' haha, right!).