Regarding 0.25um -
(EDA folks please help out here) yes it is long lead time to get designed in etc., but my suspecion is going from 0.35um to 0.25um was not easy from EDA standpoint. Need to update a lot of the tools. If that is the case, going from 0.25um to 0.18um might be a lot easier - just like going from 0.5um to 0.35um was relatively easy before.
Also, to me going to 0.25um really represents the very first generation of SOC - one just get to try out the toys so to speak, who the end customers might be, what to build, how to integrate, how to design and test etc. etc. Once those lessons are learned, going down from here to 0.18um should be a lot easier. At 0.12um, probably another hurdle in process technology, but hopefully the EDA tools and design lessons learned can be re-applied.
By the way, to suggest in 1995 that 0.25um was going to bring in money soon had got to be pre-mature. Even the leaders like Intel and IBM was nowhere near 0.25um then. I don't think one can even buy a lot of the process equipment then.
For LSI, I see them starting to generate revenue at 0.18 before mid-2000. End of 2000 latest.
patrick |
