Originally posted by: Dark Cupcake
Originally posted by: coldpower27
Originally posted by: dmens
Originally posted by: Viditor
Whatever they are, they HAVE to be better than "Core" and "Athlon", mate!
how about "core 3/4 solo/duo/quad/octo"? there, intel owes me the combined salaries of the marketing dept.
I know your joking but...
For Penryn I would go Core 2 Duo T8xxx Sequence, for Penryn 6MB and Core 2 Duo T6xxx sequnece for Penryn 3MB.
Maybe Core 3 for Nehalem derivatives.
I wonder how Intel is going to market Kentsfield, it will probably have lower GHZ then Conroe XE so how are they gonna market it against the rumored 3.2GHZ Conroe XE it's replacing? It would be a major feat if Intel had a 3.2GHZ Quad Core out, that would jsut boggle the mind.
Problem is u put that 4 core beast on the desktop board with limited fsb and ur gonna have a cpu which is bandwidth limited to the point of not being funny, and if they make it dual fsb (what i think they do in server space to conteract the limitation) than its gonna cost and arm and a leg. The HTT approach is much more clean and upgradable.
I dont know what fool desided to call it "core 2 duo" i mean thats just rediculous. Athlon and Pentium are much better names, even the k6-2 sounds better than core 2 duo. I guess they made it for all the computer iliterate people, and Amd already snatched the X2 name.
If Kentsfield is placed on a 1066FSB, I would be curious as to if it would choke Core Architecture enough that it would be a bad thing of cirtical importance.
So far from the data we have even NetBurst, which is apparently a bandwidth pig, only midly fell in performance with the switch to Dual Core on it's castrated 800FSB.
With A nomal Pentium 4 3.2GHZ is good enough to contend with the Athlon 64 3200+, on Socket 939 more or less, it takes about Pentium D 950 to be moderately competitive with the Athlon 64x2 3800+, even with the Dual Die implementation the speed didn't fall that dramatically you need about 1 or 2 higher to compensate. This is for NetBurst however.
Can you claim the same for products based on Core Architecture? High IPC Architecture, from the past leading up to this point aren't that sensitive to memory banwidth access, more memory bandwidth usually produce limited gains. Not to mention I think Intel knew that Core was going the Dual Die implementation in the near term, and designed it to limit memory access to as little as possible and not rely on FSB so much unlike NetBurst.
With Servers it's not such a big issue DIB motherboards are out now which support Dempsey with Woodcrest ones basically based on the sam tech, and Clovertown compatible ones eould basically be the same thing on updated version of the motherboards. They have the power circuitry already I believe, they were supporting Dempsey after lol, it won't be much more expensive then what we have now for motherboards for Woodcrest and even still you have to consider that AMD's solution isn't arriving till what H2 2007 at the earliest?
What I am most worried about is Tigerton, AMD has the best chance in the 4-way and up to where their stuff works to 8/16 way or so.
I will agree that AMD solution is more prefered allowing you to have access to more memory bandwidth, it's more forward looking, and seem to be the direction Intel will be heading later on as well. This doesn't mean Kentsfield is going to be strangling and gasping for air on FSB of 1.06GHZ.
Meh regarding the names, I think what Intel did is very smart, Core 2 Duo is a pretty catch phrase to me, it is just a name as long as it isn't offensive or extremely difficult to pronounce I don't really mind.