RE: Hmmm, it doesn´t come as a surprise to me either, but still, UMC was originally supposed to produce SOI Bartons even (the article states that as well)!?! It makes sense that, now that Barton is manufactured at FAB30, to go with SOI for them (which also seems to show a certain amount of confidence in it). But I still can´t understand why AMD would overestimate UMC´s capabilities as they apparently have done.
As for the rumours, it seems that quite a few were simply made up. I do think that Heye´s statements are closest to reality, but we´ll see about that.
KD2000
KD, not surprising to me either. I'm guessing that UMC is discovering what I've known for a long time. Making micrprocessor logic circuits, is a heck of allot harder than making DRAM. Many times, if a wafer yields poorly at end of line, we'll peel the chip back, layer by layer to determine root cause. Rarely, do we find something wrong in the cache areas, most of the time, the issue lies in the logic circuits. Logic circuits are much harder to make than DRAM Cache. Something that I believe UMC is beginning to realize also.
Semi