QUOTE(MmxZero @ May 30 2008, 09:14 AM)
Is your job is biz field?
Let's use life cycle costing. When the product mature, new design will need to take place in order to boost the sale. at the same time it's also cutting cost. At this stage, your cost should be already low while your sales are still constant.
When it starts to decline, the new design will come it to boost the sale in order to cover up the cost left which already inccured in the project.
Since the jasper set is shrinking the CPU & GPU to 65nm which will lower their cost. But valhala it's a new type which combine CPU and GPU. They first had to incurred cost in order to reasearch the design cost for the motherboard and whether they will solve 3rod. 2ndly the will still need to incurred cost to do research on th market whether customer will still want to but their new type of design.
Since it's new design, I doubt it will be cheap in the 1st place because it's a new technology and you are ahead in the market.
Overall, they still need to look at customer POV. a redesign in a product it's always to boost sale and to cover up cost. The only way by making the sale higher it's solving the 3rod issue which every customer POV wants.
I'm in engineering field. Not biz.
I agree with your view with design goal but not the design cost.
The cost of developing a integrated chip is negligible compared to cost of initial chip design. and the cost saved for integrating 2 chips in is beyond you imagination in large scale production....at least i don see you know it from your reply. Do remember that they still performing the same, not that they added a new functionality or design to the chip. Just have a look at Intel Core 2's example. The only big cost in this case is the new fab for new chip. but do remember that it's the fab company doing the manufacturing invesment, M$ just merely buying them at lower cost as the yield are more than previous gen.
And FYI, the problem that caused 3RROD is actually happened on all electronics--a cold welded chip,and it happened on PS3 too! it's just that xb360 interpreted them in such pattern + extremely high failure rate on previous production making a bad impression to ppl; while PS3 just died as a bricked machine and my estimation is this type of failure is less than 0.5% in PS3.
IMO, as long as the OVERALL failure is below 5% is considered a big advancement already. the 3RROD will always be there unless they removed the error mapping and/or ..... they use a quantum or DNA or other design that forbid electronics's limitation.
Hope you'll be happy with all the idea.