News Focus
News Focus
icon url

Wouter Tinus

09/18/07 1:58 PM

#48577 RE: wbmw #48569

I don't think that's possible due to the I/O pads on all sides, but I'll ask if I get the chance.
icon url

mycen

09/18/07 2:01 PM

#48578 RE: wbmw #48569

Perhaps this was done so that they could easily create a 8 core Nehalem derviative. Mirroring the 4 core and modifying the IO/CSI ports on the mirrored side could possibly constitute much of the rework for Nehalem 8 core. Hmmm.. What is Beckton again?
On the other hand, AMD's 4 core Barcelona's floormap had the L3 cache in the center and they have to go with the MCM approach to get 8 cores or to drastically re-work their design.
icon url

ChipGeek

09/18/07 2:07 PM

#48582 RE: wbmw #48569

Beamer, I'll just say this. Notice that whenever Intel mentions Nehalem, they include the term 'scalability' in the same breath. The layout of the cores plays a big part in the development of the complete product line, however as Intel isn't saying anything yet then I probably shouldn't either. Keep staring at that die, though; there are clues there for the observant...

icon url

Saturn V

09/18/07 2:27 PM

#48593 RE: wbmw #48569

Nehalem Die Picture & Four Vertical cores:

How do you scale this up to eight cores ? Stack more cores horizontally, or add four more cores at the bottom ? The second option looks better to me. Then it becomes like the AMD flag configuration.

So far Nehalem road map looks impressive for servers and data centers. CSI should improve multisocket performance, and "scalability" should cut off power in unused cores and cache,and alleviate the "global warming" due to data centers. For the client space, integrated graphics should improve performance. I hope that Gelsinger or others will give indications of the expected performance range.

Otilleni's comments at the end of Q2 earnings call made a Nehalem IDF demo a given. So I am surprised that Nehalem is only 3 weeeks old! It may not be the first pass at Silicon. Barcelona is looking very vulnerable,even when its speed path problems are fixed. AMD will have to depend on the 'Value marketing' and guerilla tactics like tricores. Profits at AMD will continue to be elusive,unless Intel commits a big blunder.

I wonder when the webcasts will be available on line ?