Easy, they will take another, more relaxed, process and release it as 10nm. Besides, technically they already have 10nm products on the market. Just don't tell anyone, that they are extremely bad
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Intel kills 10nm ?? oO
- Thread starter Zeed
- Start date
Easy, they will take another, more relaxed, process and release it as 10nm. Besides, technically they already have 10nm products on the market. Just don't tell anyone, that they are extremely bad
That rumour about being close to solving one of the hurdles was actually in SemiAccurate. (for all its Intel bashing)
But now with 3 years and counting late from first "production starts in end 15" roadmap and apparently very little light in the end of tunnel, it wouldn't be wonder if Intel decided to do overhaul of goals.
They need some kind node shrink to keep making those monolithic server chips.
So even less ambitious shrink working in foreseeable future would be better, than current status quo of no sure idea when 10nm produces even desktop chips-situation.
No need to OC anything:They are going to use Principled Technologies to OC the report
When you disable half of 14nm that makes 7nm...
It was Intel bashing in its own way though glorifying in them having to relax their design - which actually I don't think it correct - wish I understood the technical details enough to explain why - it is related to their approach to the lithography with the intention of making something on the limits of DUV which means they can't simply relax it.
The problem is less ambitious shrink needs time if kick-started from today. I believe intel has several processes worked on in parallel, though with their brain dead management, there might be a chance they cut corners
Not so much cutting corners but by all reports progress on parallel capabilities have suffered as they've (re)directed resources to try and get 10nm working.
Soldato
I'm not sure how Intel are going to be able to justify wasting billions of more dollars if they cannot sort out the shortage in their current marketplace. They are going to be bleeding market share if not careful, and the amount of wasted time and resources that have gone in to 10nm development will only get bigger, unless they either come to market with huge amount of SKU's across all their segments in the next 18 months on 10nm then they are going to suffer. Whether that be running a 10nm that is more like a 12nm, or just dumping 10nm, keeping 14nm+++++++ going, and sorting the 7nm out by 2021-2023.
Either way, the shift from 14nm to 10nm needs to happen very, very fast, or not at all in order for Intel to be seen as taking its own market seriously by the people who actually own Intel stock.
Either way, the shift from 14nm to 10nm needs to happen very, very fast, or not at all in order for Intel to be seen as taking its own market seriously by the people who actually own Intel stock.
They will justify the same way they justified spending money on failed GPU attempts, or for contra-revenued mobile chips. remember, Intel is top of the pops when it comes to spinning things
Caporegime
Soldato
Have we now reached the end point for silicon?
eee nope why would You think that ??
Maybe not of what's physically possible, but commercial viability is another story.
A number of commentators have speculated that we never get so-called "5nm" or below, purely due to the costs involved.
Associate
5nm will be affordable 3nm just but what will really show up is how much time (years) it will take to get their.
Everything gets affordable once it hits mass usage/production and new tools being developed
But when you have so few atoms to work with on the < 10 nanometer scale, the engineering challenge becomes far greater. A single silicon atom has a diameter of 0.2nm, that's only five silicon atoms per nanometer not factoring spacing. Insane!
Indeed
Look at how many foundries there were on the leading edge years ago.
Look at how many there are now.
At some point nobody will be willing to invest the sums of money needed to move forwards.
At some point therefore progress will basically just stop, because nobody will be able to recoup the investment needed.
Have to say I'm a bit pessimistic. I think we're pretty much near the apex of what we will achieve; much like manned space exploration, the will to continue just won't be there in light of the staggering sums of money involved.
"10nm" is a marketing name only.
Much of the actual gate dimensions are far larger, like 70-80nm.
Which isn't to say it's not extremely impressive, regardless. Because of course it is.
It isn't so much the money involved because money value is virtual, anyways and we can cover the costs. The problems are more political and the interests of some people.
As I said awhile back in response to that article - can't remember if it was here or another forum - a lot of what they said didn't make sense from a technical perspective when it came to 10nm and how you'd change it and as per the updated article it would be more likely they'd backport to 14nm rather than make changes to 10nm in the way claimed or move resources to 7nm.
Either way intel are screwed for a couple of years, at least. 14nm architectures won't be so competitive, while 7nm is a distant prospect.
Probably not so distant - they've been working on getting EUV upto speed with 7nm - which seems to me to have been one of the stumbling blocks with their 10nm due to pushing the boundaries in certain areas closer to what you'd see on 7nm despite other areas being more relaxed.
semiaccurate.com said: