Ashraf has been pretty negative lately, but he poses some hope culled from the recent TMG day. It's actually the first smart and rational idea any Intel CxO has offered in the last year, and it came from Moorthy, of all people.
https://www.fool.com/investing/2017/04/02/intel-corporations-exciting-plan-to-chop-up-future.aspx?yptr=yahoo
Who has the richest set of IP across multiple nodes to optimize PPAC, ain't the company spoiled by a monopoly and 60+ GM.
Murthy's fastest path to new products is on TSMC or Samsung silicon not TMG 😉
What 1-gka said, plus:
On the positive side, Intel could design custom products a bit faster which is what Murthy is all about.
On the negative side, this makes assembly even more expensive and requires much more engineering and supply chain management than Intel currently is capable of; sounds like a good idea until an excursion on $2 I/O chip is caught late and trashes thousands of assembled CPUs; gamers aren't paying for features they don't use, they are paying what Intel can get out of them for the least amount of development work and wafer real estate. If Intel cared about feature optimization the I7 gaming chips would all be high core/high cache server variants.
The sad fact is 32, 22nm and 14nm were all designed from ground up for PPA for CPU. To repurpose those more mature and higher yielding to be flexible for other things require radical new design rules, new devices and new processes developed.
TMG has been set up that they can only effectively resource one mode at a time. I would say they can't even do one node on schedule given the FUBAR on 14 and even bigger one on 10nm
They would need to change the culture and skill set in LTD. For those that know LTD and TMG they know it is a joke to believe that will happe.
OP here -
@kmb - Correct, Intel is years late. I was merely pointing out that for the first time ever, someone has figured out something critically wrong at Intel and has been public about going to obviously right direction.
@qgg - Anything I say will be met with sophomoric emojis. When do you graduate middle school?
What a revolutionary idea! I am sure Intel is the first to have thought of this and surely the first to make it work. Nobody else hs been able to commercialize this technique or patented it until now.
Oh wait, April Fools was yesterday.