Originally posted by: IHAVEAQUESTION
I thought Pentium was long dead and I didn't even know they still make Core Duo. Aren't these two designs utterly insufficient compared to Core 2 Duo? Why would anyone want them? And where does celeron come into picture now?
You're thinking is inline with a typical consumer.
In simple English it's down to business, manufacturing, marketing and best practices.
It's unfeasible to scrap all your current production from your current design to a new design since this isn't efficient to put it bluntly. To be a little more specific it's like this:
- You need time to improve yields on a new die design and the best way to find errata is by monitoring how the market reacts, you can only remove so much errata in pre-production
- You need market demand to make it worthwhile to ramp production so to not over supply a product
- You need to keep feeding the market all the products they consume and not just concentrate on your new wonderful processor as otherwise this as severe financial implications
The above brings me into manufacturing because you cannot just switch the factory to making a new die overnight. This is especially true if you're moving to a new manufacturing process which means buying new machinery for the given factory, setting it up, doing test runs, validation etc. You do not go reconfiguring your fabs to new spec until their is a demand or until it becomes financially feasible. Intel reuses previous nm machinery for its chipsets. The latest Core 2 Duo and i7 (including i5) chipsets are all 65nm, down from 90nm previously.
You must also bare in mind side issues such as the fact that Intel is obliged to keep production of various models of its processors before they become EOL. This is such that the various workstations and servers that are out there have spares. Embedded systems also need older parts since even an Atom processor is way too fast for what they do (plus it in many cases would mean re-writing the expert systems code base again).
Then that falls into marketing. It is pointless making a product that there is no desire for in the market. That's just spending marketing dollars for nothing. You need to gradually switch the mindsets of the consumers (and business users). Marketing campaigns take time, planning and lots of money. They follow a set theme and must be allowed to run their course.
All the above is also best practices as is re-using the Pentium brand where applicable. Intel has spent billions on marketing the Pentium brand since 1993 and hence using it is a valid business move, and thought through marketing decision. Ask anyone on the street if they know what a Pentium is and they will most surely say Intel. Marketing interlinks with psychology, people feel more at home with brands they recognise. Creating brands takes time, planning and money, as part of a wide series of marketing campaigns.
The Celeron has, is, and will remain Intel's economy line of CPU's for both desktop and mobile. The brand was created in 1998 and is now matured into the marketplace. The same goes for Xeon, also launched in 1998.
Once Intel's GPU is publicly unveiled, expect a new brand to accompany it too, just like we had with the Atom for 2 years+ now.
Intel no longer makes Yonah based processors. Even the slowest mobile models are all Core 2 Duo, even the Pentium branded ones albeit with less L2 cache and some featured disabled internally, and a slower FSB at that. Of course the Atom series is a whole different story.