"any idea what would be the die size of such a (256 single channel) part at 130 and 90 nm?"
About 130 mm^2 and under 80 mm^2. If they go whole hog and remove the 2 HTT ports, they could probably get it down to about 120 mm^2 or less and maybe 70-75 mm^2. They might not do that, it would make the validation more difficult and might complicate testing. Note: nothing has been mentioned about a 256k version on 90nm. It seems like a no-brainer, but Dublin/Paris isn't due out until Q3 and it's a 130nm product. I don't have any explanation for that, except maybe there are some tricks they can do to get low power that they can't do with 90nm. Assuming that Paris/Dublin is very low power. I suppose it's possible that 130nm will be yielding more die per wafer than 90nm until 2005, but...
Market Data 
Markets 
AI
