Originally posted by rickag
I doubt any significant differences exist due to ISA. I believe it ultimately boils down to die size and yields, no matter what ISA architecture.
Yup. Die size, yields, and capacity. Don't forget capacity. Intel is the leading manufacturer in 300 mm diameter wafer fabs as well. That's double the capacity per wafer compared to the 200 mm wafers dominant in the semiconductor industry.
Last I heard, Intel manufactures its chips for an average cost of about $30/chip for CPUs and probably <$10 for the support chipsets (northbridges, southbridges, etc.) Likely much cheaper than PPC 970fx and 744x chips because they sell an order of magnitude more chips. What they charge for those chips is largely dependent on the negotiating skills of the buyer.
The ZDNet blogger is quite the idiot. I mean "Woodrow?" Intel has confusing code names, yes, but come on. He's trotting out the same old misinformation all over the net and creating some new ones to boot.
The low power 970fx chips won't make fantastic laptop chips because they are limited to 1.6, maybe 1.8, GHz. Yonah will be dual-core and 2+ GHz.
The 8641D won't make a fantastic laptop chip because it won't ship in Summer 06 and will be limited to 1.8 GHz or so for laptops. Merom will ship late Summer 06 at 2+ GHz.
This statement, "Unfortunately, "Yonah", even in its first 32bit incarnation, isn't ready and its full implementation successors, "Woodrow" and "Merom", keep getting further and further behind schedule
" is pure FUD. Yonah is purely a 32 bit chip, and looks to be very on schedule for a January delivery date. Merom chips haven't fallen behind schedule whatsoever.
So, the only interesting point he makes is that he thinks Apple won't get volume discounts from Intel, or that they can't manufacture an equivalent or better Mac/Intel than a Mac/PPC which is just laughable. They can, but it's going to cost more than x86 commodity box makers. Apple's got style to charge for.