you spelled it out correctly. It is a fundamental assumption that Apple is smart enough to get a binding supply contract from samsung on price/quantity/term before starting the process. Any comments without making that assumption is non sense.
This shows complete lack of understanding what it takes to manufacture and qualify a chip like the A5. First and foremost, leading edge capacity in the amount t that sole needs it simply does not exist unless TSMC pisses off their entire customer base or Intel pr GF signs on to be an Apple foundry. Second, the A5 is far from being particularly unique, it is basically and ARM core with some Apple elements. I would not be surprised if it uses substantial physical IP from Samsung. It would take upwards of an year for Apple to secure a second source in the quantities they need. The really interesting question is what their foundry contract with Samsung says. Because I am sure that at least some within Samsung are considering whether the cost of pulling the plug on Apple manufacturing is worth it. Given Apple's aggressiveness in the trial I suspect that they have a firm contract for supply a year out and are hard at work on a transition thereafter. The really interesting piece is that there are probably only two companies that they can go to, Intel and GF. In the past Intel has walked away from non profitable Apple business, and given that Intel wants to push their own mobile chip line I suspect that GF will end up with the business.
At this point I think it is moot.
Samsung has been able to gain some traction with larger screens and a faster move to LTE. So the Korean company is now actually leading in key areas where the iPhone 5 needs to play catch up.
Blog Doing Math in FPGAs Tom Burke 2 comments For a recent project, I explored doing "real" (that is, non-integer) math on a Spartan 3 FPGA. FPGAs, by their nature, do integer math. That is, there's no floating-point ...