LinkBack URL
About LinkBacksRead more.But remains undecided on whether to use EUV technology as part of the process.
Read more.But remains undecided on whether to use EUV technology as part of the process.
a 5nm R9-690 GRAPHICS CARD? you will just need a 300w power supply nothing more! and here is the catch the R9-680 will just need Pcie motherboard bus power.
gosh we are getting close to 1nm!, what happens after we achieve this? Can we get even smaller than this?
It is fascinating to watch isn't it. 5nm is about 25 atoms across for Silicon? Not much margin for error is there!Originally Posted by NitrousX
Google tells me silicon's atomic radius is 117.6pm, so shell to shell you'd get less than 22; I guess it depends on how dense the crystal lattice is: mid twenties is certainly the right ball park. I suppose Carbon might be the next target? It is ~ 40% smaller so there'd be another node or two's leeway if they can make Carbon chips work....
Bull**** show one of these 10nm or 7nm wafers Intel hasn't revealed 10nm wafers yet so no chance TSMC is any where close.
At that scale you start getting very serious problems with quantum tunneling. 5nm as a limit is just the last 'node' on the ITRS https://en.wikipedia.org/wiki/Intern...Semiconductors because frankly we don't really know what to do at that point beyond more efficient layouts.
What will almost certainly happen is Intel et al will start dumping serious cash into technologies (like mass production of graphene) that currently aren't affordable at scale until they find one that works.
Personally I'd quite like it if someone said that that's the end of the line for a couple of years because it'd kick software developers into writing code that isn't so bloated (because we can afford to be).
Last edited by Whiternoise; 18-12-2015 at 04:25 AM.
"Starts work on" is just the very beginning of a long process, that is as much news as "I wish I was on summer holiday, though I haven't even decided where I want to go this year let alone booked it".
Samsung are far enough along on 10nm that they have been showing test wafers: http://semiaccurate.com/2015/11/12/s...afers-techcon/ so I guess that is where everyone is at. Intel are saying they won't be using silicon at 7nm (http://arstechnica.com/gadgets/2015/...ilicon-at-7nm/), perhaps that is why they seem to losing momentum at the moment because if they can pull off a technology jump like that it would be very impressive.
Word. It's good to know they have solid plans to go further but the fact that they haven't even settled on lithography tech yet says rather a lot...!
Even now electrons can go AWOL (quantum tunnelling); the narrower the matter keeping them on path, the more likely it can happen; hence the move to other materials is being considered. Basically silicon is coming to the end of its life for cutting edge tech; doesn't mean it will cease over night, or even any time not-so-soon.
I was told that when I did my electronic engineering degree, and that was over 20 years ago. That's the thing with a frontier of technology, we have always had one
There are currently 1 users browsing this thread. (0 members and 1 guests)
Posting Permissions
Reply With Quote