Does AMD have an answer to Nehalem?

Everyone is kind of expecting the jump from Core arch to Nehalem to be as big as the jump from Nutburst to Core.

I think everyone needs to remember that Core arch was revolutionary as it took the mobile CPUs and adapted them to the desktop market.

Nehalem is just an evolutionary step in the Core arch. i would expect it to be similar to the increase seen going from Pentium to Pentium 2, not like Pentium D to Core 2 Duo.

I don't think AMD is as doomed as everyone thinks.
 
Everyone is kind of expecting the jump from Core arch to Nehalem to be as big as the jump from Nutburst to Core.

I think everyone needs to remember that Core arch was revolutionary as it took the mobile CPUs and adapted them to the desktop market.

Nehalem is just an evolutionary step in the Core arch. i would expect it to be similar to the increase seen going from Pentium to Pentium 2, not like Pentium D to Core 2 Duo.

I don't think AMD is as doomed as everyone thinks.

I think the jump is going to be very much like the jump from the Athlon XP to the Athlon 64. Not every application will benefit from the change, but I'm betting games certainly will.
 
Unlike R700, I doubt Deneb will be near the ballpark performance wise (its only equal to Kentsfield per clock), so effectively AMD is competing with an architecture that is a generation ahead.

Not to mention, a 2.66GHz Nehalem costs $284, what kind of trick do you see AMD pulling? Sub $200 Denebs? What about margins?

I'm not saying that Deneb would be the answer. Deneb would probably be like HD3800, a die shrink of R600 to learn more about the petential of their architecture. When they know the strength and weakness of their architecture, they could then improve it like how they did with R700.

People were saying that 16 ROPs were the bottleneck of their R600 architecture and the ROPs were the reason why the AA performance was so bad but now they managed to get an insane AA performance while keeping the same number of ROPs like before. They have managed to shut the people up now.

I hope that they manage to pull the same trick with their next CPU architecture. Looking at their situation now, I think that I have a rough idea on the direction they are going in the future.
 
Everyone is kind of expecting the jump from Core arch to Nehalem to be as big as the jump from Nutburst to Core.

I think everyone needs to remember that Core arch was revolutionary as it took the mobile CPUs and adapted them to the desktop market.

Nehalem is just an evolutionary step in the Core arch. i would expect it to be similar to the increase seen going from Pentium to Pentium 2, not like Pentium D to Core 2 Duo.

I don't think AMD is as doomed as everyone thinks.

http://www.anandtech.com/cpuchipsets/intel/showdoc.aspx?i=3326
 
nVidia spent a LOT of money on R&D for GT200 too. Hmm...

ATI and NVIDIA are closer in size and finances than Intel and AMD are. Intel spends something like 4 billion dollars a year on R&D.
 
one thing could be true if amd is unable to compete intel can ten raise prices which in the end mean bad for us consumers, so who knows maybe bulldozer can do something but in the end competition is all good
 
I'm not saying that Deneb would be the answer. Deneb would probably be like HD3800, a die shrink of R600 to learn more about the petential of their architecture. When they know the strength and weakness of their architecture, they could then improve it like how they did with R700.

People were saying that 16 ROPs were the bottleneck of their R600 architecture and the ROPs were the reason why the AA performance was so bad but now they managed to get an insane AA performance while keeping the same number of ROPs like before. They have managed to shut the people up now.

I hope that they manage to pull the same trick with their next CPU architecture. Looking at their situation now, I think that I have a rough idea on the direction they are going in the future.
why would people say that 16 ROPs were the problem on R6xx when AA was done in the shaders anyway? now with R7xx the AA is not done in the shaders which was the problem in the first place not the 16 ROPs. anyway I hope they get their shit together soon on the cpu side of things.
 
I also believe that AMD has an answer for Nehalem but it would be available sometime 2009. They've pirated me to work for them in Singapore hehe. Can't wait to start working for the underdog.
 
Well the truth is that the TLB bug didn't really matter all that much. It was just a PR disaster for AMD and that's all. I think Intel is moving so fast in order to keep AMD off balance. Intel's R&D budget is probably larger than the amount of cash AMD sees all year. Intel is leveraging their pockets in order to keep the processor development ahead of AMD. This means that in order for AMD to leap ahead they'll have to get lucky.

the tlb bug was a disaster it put amd 6+ months behind cuase they had to fix the bug and then shrink the chip for 45nm this is why stuff is so late with amd right now you pull back 6 months ago amd should of had shanghai out and be working on the bulldozer core for release next year remember and does not have intels resources
 
the tlb bug was a disaster it put amd 6+ months behind cuase they had to fix the bug and then shrink the chip for 45nm this is why stuff is so late with amd right now

They were more than six months behind Intel to begin with.
 
amd will just compete in mid range in the mean time. I'm an AMD fan, there's no way around how Intel is executing and very impressed by their last two years of performance.
 
amd will just compete in mid range in the mean time. I'm an AMD fan, there's no way around how Intel is executing and very impressed by their last two years of performance.
The problem is that AMD is stuck squarely in the bargain basement category right now. Have you seen their ASPs? Until this changes, AMD is going to continue losing hundreds of millions of dollars every quarter.
 
in these days;amd had a presentation in Serbia.and they say new 45 nm cpu can be overclocked up to 4 ghz(+) speeds with stock fan:)and what will happen with henalem to intel?will they be able to offer us high speed cpu s after integrating the memory controller in to cpu?
this is will be a gamble for amd.they will screw up;or take a deep breath to go on their way...i hope they success it:)
 
Back
Top