something seriously f*cked with the 90nm process

[boran]

There seems to be quite an issue with that new 90nm process as it seems now a 130nm proc is going to beat any 90nm proc any day in wattage for mhz ... anyone have any insights as to why this is happening ?

 

[Sid59]

"It seems like AMD and Intel made an agreement. Intel went to AMD and said, "Just for a change of pace, instead of trying to outdo each other, let's see who can f--- up more this year."

AMD agreed, figuring it had a lock on that, then Intel blindsided them with Prescott. AMD said, "This is going to be pretty hard for even us to beat. Hmmmm. Until we can screw up our 90nm process, why don't we just price the s--- out of our Hammers so nobody will buy them, and obsolete the ones we have out there?" "


--- that was pretty funny.

 

[orion7144]

It's normal for a new process to start off like this. It will improve to be alot better than the 130nm it will just take time. Unfortunately it will probably take longer than the 130nm since the line widths are so small it requires much more process control.

 

[Accord99]

Originally posted by: boran
There seems to be quite an issue with that new 90nm process as it seems now a 130nm proc is going to beat any 90nm proc any day in wattage for mhz ... anyone have any insights as to why this is happening ?

A lot of the problems seems to be that Prescott has so many more transistors than Northwood. Williamette->Northwood went from 42M transistors to 55M. Prescott has 125M. If Intel did a simple shrink and doubled the caches, it would have resulted in a 80-85M transistor core and would be much cooler. The big question is why Intel went with Prescott and what hidden features are currently disabled on Prescott. Johan of Aces Hardware also mentioned that Intel promised cooler runner models in the future.

And there's also the problem of leakage which grows rapidly as processes shrink, and will impact all manufacturers of bleeding edge processors. Only future technologies like Fully-Depleted SOI, multigate transistors and high-K dielectrics will be able to manage it.

 

[orion7144]

Originally posted by: Accord99

Originally posted by: boran
There seems to be quite an issue with that new 90nm process as it seems now a 130nm proc is going to beat any 90nm proc any day in wattage for mhz ... anyone have any insights as to why this is happening ?

A lot of the problems seems to be that Prescott has so many more transistors than Northwood. Williamette->Northwood went from 42M transistors to 55M. Prescott has 125M. If Intel did a simple shrink and doubled the caches, it would have resulted in a 80-85M transistor core and would be much cooler. The big question is why Intel went with Prescott and what hidden features are currently disabled on Prescott. Johan of Aces Hardware also mentioned that Intel promised cooler runner models in the future.

And there's also the problem of leakage which grows rapidly as processes shrink, and will impact all manufacturers of bleeding edge processors. Only future technologies like Fully-Depleted SOI, multigate transistors and high-K dielectrics will be able to manage it.

Just currious, did you mean Low-K. When I was starting up an R&D Fab in 01 the big push was for the Low-K Dielectric. Also don't forget that they have already gotten the NannoTube technology working
Seen here which should help with the leakage current but is probably a few years away from integration into production.

 

[Accord99]

Originally posted by: orion7144
Just currious, did you mean Low-K. When I was starting up an R&D Fab in 01 the big push was for the Low-K Dielectric. Also don't forget that they have already gotten the NannoTube technology working
Seen here which should help with the leakage current but is probably a few years away from integration into production.

High-k:

http://www.intel.com/labs/features/si11031.htm

Thanks for the link, a lot of interesting new technologies but the unfortunate thing is they all seem to be a few years away.

 

[Mday]

speaking of K, i am reminded of copper interconnect arguments made a while back... amd was doing research on copper interconnects due to the limitation of aluminum. while intel was going on about dielectrics... hehee