On Sat, Dec 8, 2012 at 4:58 AM, Alan Grimes <[email protected]> wrote: > I'm pretty sure we're going to the 7nm node from where we are now. (we're at > around 22-26 nm as intel puts AMD in a choke-hold... we could stay at that > node for decades if AMD goes under though. =(
That is one thing that could go wrong. Another thing that could go wrong is this: nowadays the most important figure of merit is performance per watt. What happens if we reach 15nm and it turns out that because of leakage current or whatever, it actually takes more energy for a given level of performance than 22nm? I'm not saying that will happen, of course - I'm saying nobody knows whether it will happen or not. Of course even in that scenario there would still be room to improve performance per watt _if_ we could trade away ease of programming with a given level of software tools technology - which in practice means if we could improve software tools technology to keep ease of programming roughly constant - or as you say, > What we really need to do is get serious about the software. =| I'm working on that but I can't in all honesty make more than wild guesses about schedule and I have to admit history is littered with interesting hardware architectures that got to the point of shipping silicon then died because the software tools didn't materialize. ------------------------------------------- AGI Archives: https://www.listbox.com/member/archive/303/=now RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424 Modify Your Subscription: https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657 Powered by Listbox: http://www.listbox.com
