Intel kills off the 10nm process - For real??????

[lexluthermiester]

If we're doing history lessons, Core was an evolution of the Pentium-M at the time. Would you venture a guess about where their next answer to AMD will be coming from?

Honestly, my theory above about a new substrate material is the best I have to offer. It's well known that most IC makers have been looking for an advancement in that area and it's likely to be the answer to the problem of Moore's law(which is less a law and more a theory). Beyond that I have no idea.

And looking at things from a slightly different perspective, a rushed response followed by a more adequate one is what you always do when you get caught with your pants down. I mean, when Core hit, AMD had to regurgitate Bulldozer designs till they got to Zen. It's a pattern.

You're thinking of K10, Bulldozer didn't come along until 2011, but your point is valid. This pattern of advancements has been happening for decades and will likely continue.

 

[R0H1T]

Very doubtful. Seems like fear mongering created by well placed fanboys. Intel is working on an answer to Ryzen. They're as likely to give up on 10nm as the Sun is likely to stop shining tomorrow.

Charlie at SA is more accurate & knowledgeable than your avg fanboy, not to mention he has his sources well placed & well informed within the industry.

Right & I know which of these is happening 100% tomorrow, the other one has at best 50/50 odds.

 

[lexluthermiester]

Charlie at SA is more accurate & knowledgeable than your avg fanboy, not to mention he has his sources well placed & well informed within the industry.

Maybe that's true, but it doesn't mean they're as accurate as they need to be to be taken seriously. And as Intel has been making strides in many areas of the industry, the idea that they're giving up or slowing down is a silly one that lacks credibility.. It would be like saying AMD has hit a brick wall with Ryzen not being able to get beyond 4.2ghz stable and is giving up. It's just nonsense. Both companies will forge ahead and continue advancement.

 

[R0H1T]

Maybe that's true, but it doesn't mean they're as accurate as they need to be to be taken seriously. And as Intel has been making strides in many areas of the industry, the idea that they're giving up or slowing down is a silly one that lacks credibility..

They're going for 7nm, presumably with EUV. So they're not stopping for sure, slowing down compared to their own pace of development years ago. They also have enough market share & profits to weather this particular storm, having said that even at 7nm they'll need something of a Hail Mary to turn their fumbling fab business around.

 

[Vya Domus]

I have a hard time to believe this as well. If true, there is no other way to put it : they fucked up big time.

 

[lexluthermiester]

to turn their fumbling fab business around.

It's hardly fumbling when they're expanding production due to having a hard time keeping up with sales demands;
https://www.techpowerup.com/forums/...ireland-to-increase-processor-volumes.248764/

 

[R0H1T]

It's hardly fumbling when they're having a hard time keeping up with sales demands;
https://www.techpowerup.com/forums/...ireland-to-increase-processor-volumes.248764/

Selling to your established base isn't the same as delaying 14nm by 2(?) years or 10nm by 4(?) & I did mention fab, only that part. Intel is much bigger & diverse than just the fabs they own.

 

[lexluthermiester]

14nm by 2(?) years or 10nm by 4(?)

Did you mean 10nm & 7nm? They're already making 14nm and have been for years.

 

[StrayKAT]

I'm sure if some outlet said Intel is secretly just some flea market operation based in a ghetto neighborhood, some people would actually believe it.

 

[R0H1T]

Did you mean 10nm & 7nm? They're already making 14nm and have been for years.

14nm was also late, IIRC by a year if not two & 10nm is still on a sabbatical.
A year it is, two if you count mass availability of 14nm chips in 2015. Surely you remember Tic Toc ~ https://en.wikichip.org/wiki/intel/process

 

[lexluthermiester]

14nm was also late, IIRC by a year if not two & 10nm is still on a sabbatical. A year it is, two if you count mass availability of 14nm chips in 2015. Surely you remember Tic Toc ~ https://en.wikichip.org/wiki/intel/process

Even if it's just next year, a delay is not giving up. It's just a delay.

 

[Papahyooie]

Or they are renaming it to 7nm, which makes sense since 10nm Intel process is on par with TSMC's 7nm. Better marketing

Honest question, how could they get away with something like that if that were the case? I was under the impression that the node rating is an actual measurement of the gate size. If the gate size is 10nm, you can't simply call it 7nm. I mean I guess they could just lie... but fab process isn't "marketing." It's an actual measurement. I'm sure there would be some legal implications if they did that.

And what does "10nm Intel process is on par with TSMC's 7nm" mean? Either the gate width is 10nm or 7nm. You can't say that 10nm is on par with 7 nm. It doesn't make sense.

 

[R0H1T]

Even if it's just next year, a delay is not giving up. It's just a delay.

If 10nm is still on track, no one apart from Intel Inside(r) knows this for certain ~ unless we take Charlie's words at face value.

 

[lexluthermiester]

unless we take Charlie's words at face value.

Which we can't, so the whole debate is superfluous.

 

[Frick]

Even if it's just next year, a delay is not giving up. It's just a delay.

Honestly it's nothing "just" about it at this point. They should be nervous. The strain is definitely showing.

 

[R0H1T]

Which we can't, so the whole debate is superfluous.

Do you, for instance take Videocardz leaks seriously? I bet their hit rate is just about the same what Charlie delivers, that is to say in terms of speculation this is no better or worse.

 

[Sasqui]

Made me laugh a little. Reminds me of the Pentium 4 days. That design was stagnating so bad the only thing they could do was to add a second die and call it a dual core. While that worked it was just a stop-gap solution until the Core/Core2 design released. The last 18 months are mirroring those same events. AMD forges ahead with a great performance lineup, and Intel answers later with a breakthrough. History seems to be repeating with a few interesting changes. Intel is working on an answer to Ryzen. You can bet your life on it. How good an answer it will be, who knows..

That brings back memories. P4 sucked ass, then all of a sudden people were staggered by the intel laptop chips that seemed to blow the P4's away! Not long after, the Core desktop chips came out, based on that architecture. And Intel was back in the game again.

AMD had the on-die memory controller that still kept them in the game... until the i7 came out.

Now Ryzen, and intel is feverishly working on an answer. 10nm is a fabrication problem and it will be solved, who knows when, and like you say, who knows how good it'll really be?

At the end of the day, you've got to love competition!

 

[lexluthermiester]

Do you, for instance take Videocardz leaks seriously? I bet their hit rate is just about the same what Charlie delivers, that is to say in terms of speculation this is no better or worse.

Not really. Rumors are just rumors and not to be taken seriously until supported with verifiable evidence. Rumors can be fun & interesting to read about and also can be fun to debate as long as people don't get angry about it.

The strain is definitely showing.

No it isn't. Intel is experiencing very high sales and demand. They're not at all strained with anything other than supply shortages.

 

[Vayra86]

Not really. Rumors are just rumors and not to be taken seriously until supported with verifiable evidence. Rumors can be fun & interesting to read about and also can be fun to debate as long as people don't get angry about it.


No it isn't. Intel is experiencing very high sales and demand. They're not at all strained with anything other than supply shortages.

It is always wise to see what a rumor is really made of. And having read the SemiAccurate piece front to back, there are very plausible arguments put forward and echoed over the past few years. If you look at the time line of quotes they put at the beginning of the article, and I know all of them from when they were hot news, it falls into place nicely.

We also know recent Intel changes such as moving people around, readjusting their fab budgets etc. You don't publish such major shifts because '10nm is on track'. You do it because you're adjusting to a new reality - in other words - a change at the strategic level.

 

[TheoneandonlyMrK]

It's hardly fumbling when they're expanding production due to having a hard time keeping up with sales demands;
https://www.techpowerup.com/forums/...ireland-to-increase-processor-volumes.248764/

They're struggling to replace all those ,now unwanted server chips that have spectre and co flaws... .
Or do you believe Aws Google and co are fine with Intel's reassurances.
And
They're not working on graphene or arsenic based chips yet because they are theoretically a swap for silicon , graphenes bandgap issues require more work yet and the others just in material research labs atm , neither has anything in place yet to make all the things a chip needs , workable bandgap, or any components designed and verified.

Chip's are made with science not just dave and john messing about in a workshop out back, they're year's away , hence why 5nm is being worked on, silicon is the thing for at least another 7 year's probably 10.

@Vayra86 Bang on , exactly what I should have said.

 

[lexluthermiester]

Except this just happened;

https://twitter.com/i/web/status/1054397715071651841

So, these untrue rumors turn out to be just that and Mr. Charlie was just making a lackluster theory based on poor information.

 

[bug]

Except this just happened;

https://twitter.com/i/web/status/1054397715071651841

So, these untrue rumors turn out to be just that and Mr. Charlie was just making a lackluster theory based on poor information.

Even if the rumors were true, what difference would that have made for people over here? It's not like anyone here bet money Intel's 10nm will fail.

 

[bonehead123]

History seems to be repeating with a few interesting changes

"History teaches us that History teaches us nuthin !"

Gen George S. Patton

 

[R0H1T]

Except this just happened;

https://twitter.com/i/web/status/1054397715071651841

So, these untrue rumors turn out to be just that and Mr. Charlie was just making a lackluster theory based on poor information.

Untrue because Intel said so? Nope, they'll be debunked only when Intel mass produces 10nm chips ~ which is still a ways off.

 

[TheoneandonlyMrK]

Except this just happened;

https://twitter.com/i/web/status/1054397715071651841

So, these untrue rumors turn out to be just that and Mr. Charlie was just making a lackluster theory based on poor information.

And what did intel Say every other time charlie called them out about their Rock solid 10nm plans??

They do have 10nm in some laptops, with discrete graphics yet no desktop and now rumours it's dead.

10nm isn't and never will now be ,economically viable for intel ,for some reason the shit yeilds they Are getting are not only so bad it won't ramp right(3years) but the designs Are Not working as intended (Gpu dead is the assumption based on smoke and mirrors).

That's not a workable situation and if they don't ditch it and move on they're going to get truly left behind by other foundries namely tsmc , Samsung shit even glofo has tech now intel would love to have ,all while they're working out the Wrinkles on 10nm.

Intel are going to be fine, for a bit but not with 10nm ,all my opinions on it though.

But on your opinion of Charlie and his site, wow you don't visit it do you, it's piss poor, hardly ever a new post ,first in a week ,that one ,im not paying his fee though so might be busy i wouldn't know, point being ,it is not for clicks , and if you read it the posts he lists he Was right.