Elmer -
"Was RDRAM manufacturable?"
Sure.
A bit of revisionism? Manufacturing problems plagued the technology for at least two years, enough time for DDR to overcome its late start and take over the market. I'm not even counting the disasterous MTH, which was Intel's attempt to buy time for the ever-late unmanufacturable RDRAM against the tide of DDR. By the time it was manufacturable, the market window was closed.
"How about the huge-die Itanium?"
What about it?
Way late to market, produced only in limited quantities through the first two generations, still very expensive for the mass market. This is the very poster child for experimental technology that is expensive to manufacture - due to the huge cache required. Doesn't exactly fit with your statement Intel could easily design the highest performance product in the world if they didn't care if it was manufacturable or could produce a profit. They did try to produce the highest performance product, smug in the believe that no one would challenge their x86 leadership (forcing a high performance x86 race), and spent billions in doing it.
"Soon we will see about Prescott - the rumors are in full agreement that it is difficult to manufacture."
Ahhhh the "rumors"! I'd like to comment here but can't.
Of course. No need to comment, the first parts will (finally) be here soon enough.
As I do not wish to play ping-pong, you may now have the last word. :)
AI
Market Data 
Markets 
