What's the bet on Willamette?
Delayed or just Coppermined? We spill the beanies
A couple of hardware sites have said Willamette is late, based on a hesitation senior Intel VP Paul Otellini made at a conference call earlier this week. Someone asked him if Willamette products would start arriving in the year 2000, and then he corrected himself and said, no, the year 2001. And at the AMD conference call, execs were saying that the K7 Athlon would have little or no competition next year as Willamette wouldn't arrive. So we rang our usual chums at Intel who would merely say that Willamette is scheduled to arrive in the 2000-2001 timeframe... But this is what we already know about the Willamette platform. Chips could have as many as 450 pins in its socket design, use dual channel Rambus memory and employ a chipset called Tehama. Willamette, as well as Merced, have existed as design concepts for seven years so far. Merced was then called P7 and Willamette P68. Willamette will have RISC-like storing trace caches and Coppermine and its derivatives will likely be the last of the P6 line." There could be as much as 128K level one data cache on Willamette processors. The cache and the core of Willamette architecture are totally different from P6 technology. The L2 cache, which is likely to be 1Mb, is broken into hundreds of squares, suggesting a massively parallel schema. As for Coppermine itself, Intel will release a 666MHz CuMine chip in November, its problem being that it cannot, yet, exceed 550MHz. As our friends at the Microprocessor Forum suggest, this is likely to cause more problems with Intel's mobile CPU platform than the desktop platform. We've been urged to write the Willamette "is late" story and would dearly love to do so. But as we don't have a date for Willamette, even though Paul Otellini said it was "on target" at last February's Intel Developer Forum, it's pretty hard to do that. If it is late, then we think the reasons are likely to be marketing, rather than technical. Intel will want to push Coppermine for all its worth next year before taking the next step. All these roadmaps certainly make for a complicated existence. ®
Sponsored: Benefits from the lessons learned in HPC