dmf, Intel didn't put the "A" engineering team on the initial Rambus evaluation. They put the "D" team on it (or none at all). In any event, management strongly committed to it, started implementing chipsets with it, and the many problems surfaced. All the while, they found no easy way to get out of the embarassing situation. Finally, the shaky Rambus interface (both sides, Intel and Rambus, who knows who was most at fault), high price of Rambus DRAM, strong resistance to adopt Rambus by the DRAM makers of the world, and dislike of Rambus by Intel's customers, and the OEMs, forced Intel to come out saying Rambus is not the greatest thing since sliced bread. Personally, I think the "Rambus advantage" is way overblown, and definitely not worth the headaches, both in the technology, and politically with the DRAM vendors and the OEMs. The advances in cache memory and impending DRAM on chip make Rambus a don't care. Intel doesn't need them.
Tony |
