> > > > Yeah I think that was the point, it meant that they COULD > purchase the > > cache if they wanted which freed up production capacity for > the core. > > This then drove down costs as they would have been able to get far > > more actual processors on a wafer than if it was > integrated. Also at > > the time the packages that were being used in the industry > were very > > expensive for larger dies compared to the ones we are using now. > > > > It may very well be that the marginal cost for a larger die > was greater then > than it is now (largely due to the move to 12" wafers, I > imagine), but let > us not undervalue just how huge the PPro core was, even > compared to today's > dual core chips. > > The 350nm PPro with 512KB cache was 438 sq.mm, vs. the largest chips > available today: the Pentium D dual core at 206 sq.mm and the > X2 Toledo > (2x1MB) at 199 sq.mm. > > The PPro was an extremely costly-to-manufacture beast, and > still would be > today. > > Greg >
Absolutely. I wish I still had some of my old copies of Byte magazine (it's a real shame they stopped printing it, I loved it). I remember when the PPro was the cover story, I was at school at the time and a few of us were huddles round Byte amazed at the images, specs and technical breakdown they did of it. :)
