backtop


Print 53 comment(s) - last by RMSe17.. on Sep 26 at 11:56 PM

Four 2.66GHz Intel CPUs on a single package for just a grand

Intel's newest roadmap started making the rounds last week, and the headline title across the roadmap was absolutely Kentsfield, Intel's upcoming quad-core processor.  During last month's investor call, Intel CEO Paul Otellini announced the chips would be shipping this year, as opposed to Q1'07 that was originally slated on the roadmap.

Now a few additional details of Kentsfield have slipped out.  Intel's most recent roadmap claims Kentsfield, which will ship as a Core 2 Extreme branded processor, will run each core at 2.66GHz and a 1066MHz front-side bus.  Essentially, the processor is two Core 2 Duo E6700 processors packaged onto a single CPU.

There is no announced ship date of Kentsfield yet, though Intel has announced that the processor will ship for $999 -- the same as every other "Extreme" processor the company has announced.  Intel has no price cuts for the E6700 processor planned until after the quad-core Kentsfield launch as well.  Since the E6700 has a distributor price of $530, the Kentsfield actually offers some discount for the second core.

Absentees from Intel's roadmaps include a 3.2GHz Core 2 Extreme processor.  The company officially stated that such a processor would follow the existing 2.93GHz Core 2 Extreme that is available today.  However, it seems fairly evident that the quad-core Kentsfield has
supplanted any new dual-core Extreme processors.


Comments     Threshold


This article is over a month old, voting and posting comments is disabled

RE: Fast Forward?
By Targon on 8/17/2006 1:18:07 PM , Rating: 3
I think that you may be misunderstanding SOME members of the crowd that point out how AMD has a better design. There are several issues going on right now when it comes to AMD vs. Intel in terms of performance.

Intel has a huge advantage over AMD with their 65nm fab process right now, and as a result, the clock rates for the Core 2 processors are up there. When AMD also has their 65nm process ready even the current design of the Athlon 64 without any other change would allow AMD to catch up in terms of performance.

As you increase the number of cores, a better system architecture will allow the performance of AMD processors to catch up.

With the integrated memory controller on current generation AMD processors, going to 1066MHz DDR 2 memory will seriously reduce the lead that Intel has, even with nothing else. Again, as the memory speed increases, AMD gains more system performance than Intel does. This implies that once DDR-3 comes out, even without other changes(besides the memory controller), AMD may catch up to Intel in terms of performance.

That's why AMD fanboi types feel that AMD isn't in trouble yet, because if you add up 65nm, K8L design improvements, and improvements to memory speed, AMD won't be behind for as long as people think. Intel still hasn't come up with a new overall system design for ages now.

With AMD pushing for a tighter connection between CPU and GPU(the AMD/ATI merger being evidence of this), there may also be more of a split between the system design of an AMD based system and Intel based system in the future. HyperTransport really changed how an AMD based system works on the motherboard, and going forward, AMD seems to be looking for yet more ways to improve performance on a system level. If/when HTX becomes used for graphics cards, an AMD based system may have yet another advantage over PCI-Express based Intel machines.


RE: Fast Forward?
By Motley on 8/17/2006 1:49:43 PM , Rating: 2
You seem to miss the point that even with the 65nm die shrink, and the faster RAM, AMD's performance for the Athlon will only increase marginally (8-10%). It's no where near enough to even catch up, and that all assumes that Intel sits around and does nothing for the next year, year and a half while AMD tries to get all their ducks in a row.

As AMD moves to 65nm, Intel will most likely have already moved to a 40nm process. Nothing changes there, they both can release a slightly faster CPU (A marginally faster one). The fact that the conroes overclock so well shows that Intel COULD if they wanted to release a MUCH faster chip right now if they wanted to. They would just have to speed test and bin, and they would have PLENTY.


RE: Fast Forward?
By Ringold on 8/17/2006 5:31:07 PM , Rating: 2
quote:
ntel sits around and does nothing for the next year, year and a half


I figure, except for speed bumps, they will, just like they did with the P4 platform in general.

I think the bottom line guys like Targon would want to get across is at least that Intel doesn't have the absolute advantage. Intel fans are currently 'right', just like a broken clock is right twice a day, and are doing everything they can to relish it, but if K8L and 65nm speed-boosts really really come to bear some sweet fruit AMD will be right back in the game and, as almost always, with a lower overall platform cost or price/performance ratio.

And besides, elegance counts for something. If you dropped a V12 in a crapy early-90s Civic and it could do jaw-dropping quarter miles, I'd still take a sleek, elegant but slower competitor. :)


RE: Fast Forward?
By zsdersw on 8/17/2006 8:29:03 PM , Rating: 2
I wonder if you'd be saying the same about "elegance" if/when AMD released a two-cores-on-one CPU.


RE: Fast Forward?
By Tom Tom on 8/18/2006 9:23:08 AM , Rating: 2
Where did you get your 8-10% increase figures from? The engineers at AMD have stated they expect a 40% increase in transistor performance from the shrink to 65nm. In addition some cpus are going to see a 60% performance per watt increse in 07.
"As AMD moves to 65nm, Intel will most likely have already moved to a 40nm process"
They better be ramping their 45nm right now cause AMDs ramping 65 now and shipping by Q4. AMD also stated they are going to 45nm in 08, only months behind Intel.


RE: Fast Forward?
By zsdersw on 8/18/2006 1:39:19 PM , Rating: 2
An increase in transistor performance of 40% does not necessarily translate into an overall performance bump of anywhere near 40%.


RE: Fast Forward?
By Tom Tom on 8/19/2006 8:54:03 PM , Rating: 2
"An increase in transistor performance of 40% does not necessarily translate into an overall performance bump of anywhere near 40%."

I didnt say it did, but I did say they claim they will see a 60% performance per Watt increase.




RE: Fast Forward?
By zsdersw on 8/17/2006 5:11:48 PM , Rating: 2
Increases in memory bandwidth are not necessary for AMD's chips at this point. Switching to a faster memory type isn't going to help much. AMD's chips already have a huge memory bandwidth advantage over Intel's Conroe and Woodcrest chips.. and yet aren't better performing.


RE: Fast Forward?
By zsdersw on 8/17/2006 8:31:02 PM , Rating: 2
One more thing about Targon.. he's the same AMD fan who said that Conroe's speed advantage over AMD chips is largely due to the 4MB L2 cache. As results have shown, that's a completely bogus thing to say. As such, I'm taking everything he's said with a grain of salt.


RE: Fast Forward?
By zsdersw on 8/17/2006 8:33:23 PM , Rating: 2
quote:
Intel still hasn't come up with a new overall system design for ages now.


CSI isn't all that far off.. in terms of nearly everything coming down the pipe from AMD.


"Spreading the rumors, it's very easy because the people who write about Apple want that story, and you can claim its credible because you spoke to someone at Apple." -- Investment guru Jim Cramer

Related Articles













botimage
Copyright 2014 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki