Kap, Re: "Another funny thing is that Intel, after ridiculing SOI for years, now saying that they only ridiculed PD SOI. According to Intel FD SOI is the "correct" SOI, because it lacks the floating-body effect."
As funny as it may sound, that is their stance. And IMO, it seems pretty reasonable. They've always said that they have reviewed SOI, and found it unfit for multiple process generations. AMD comes along and says that SOI is great, and they can't wait to use it right away, yet Intel remains steadfast. Later, Intel has an SOI of their own, and everybody wants to cry "hypocrite". Of course, the simple explanation is that they just happened to find an SOI process that worked for them.
Re: "Now it looks like the floating-body effect can be used to build super dense and super fast caches. I wonder if they will eat their words again soon?"
Obviously, having the floating body effect in a CPU is still a bad thing, Kap. There is still a trade-off that will have to occur between dealing with the floating body problems in the core logic, and benefiting from the advantages of having a fast, dense cache. Of course, things are still open to someone finding a way for even CPU logic to benefit, and maybe then Intel would have to eat their words.
wbmw |