Well i cannot see the intel post, but that's a good move to swiftly debunk the thing. They have just reorganized the unit and changed management... i guess the employees of the unit dedicated to the process were very anxious with that rumor.
-
Ever wanted an RSS feed of all your favorite gaming news sites? Go check out our new Gaming Headlines feed! Read more about it here.
-
We have made minor adjustments to how the search bar works on ResetEra. You can read about the changes here.
Rumor: Intel Killing Off 10nm Process
- Thread starter Vanta Aurelius
- Start date
Well i cannot see the intel post, but that's a good move to swiftly debunk the thing. They have just reorganized the unit and changed management... i guess the employees of the unit dedicated to the process were very anxious with that rumor.
Interesting. And let's be honest, these numbers are so meaningless (in a lot of ways), they could still claim it was 10nm when it actually got around to shipping. It's not like they could be accused of lying, since it's all arbitrary anyways.
Nobody would be surprised if they'd actually cancel their 10nm process at this point. Thus the debunking since the rumor is plausible enough to affect their stocks if not sales.
It's not true according to what I've heard btw. But it's hard to say how a 10nm process which will eventually be used for Ice Lake will compare to what Intel were planning to use for Cannon Lake back in 2016. There's a huge chance that they will be significantly different.
OP
OP
Do you think Intel is retooling to have a process structure similar to TSMC with their 7nm process? Or, is Intel just going to make further enhancements to their 14nm process and call it 10nm?
I'm just not sure of what scrapping an entire node means when Intel is dependent on launching a new node dubbed with 10nm. Are they pretty much succeeding the performance crown to AMD and the node advantage to TSMC and Samsung? Will Intel be able to regain their advantage?
This is not how research works. Everything that's put in the "research" section of this image cannot technically be late; those dates are just tentative guesses for which they had to be optimistic to look good. But you can't really talk about delay for something that is in a research phase at the time this image was made.
Intel's 14nm is not the same as other foundries' 14nm.
AFAIK, progress is better on their future smaller nodes. 10nm ended up being a nonstarter for them for various reasons (they say they're still working on it, but I'm not sure anybody really knows why).
OP
OP
Gamer's Nexus has a great video that goes into the possibilities around what issues Intel has ran into.
It looks like one reason things aren't working out was that Intel's jump to quad-patterning might have been too costly for 10nm to be economically viably.
I think a lot of the hardcore PC gamers on this forum already know this, but people shouldn't read a lot into the node naming conventions. TSMC '7nm' will probably comparable to Intel's 10nm. Node process names stopped meaning anything a long time ago.
https://en.wikichip.org/wiki/technology_node
hope this helps with the confusion.
https://en.wikichip.org/wiki/technology_node
hope this helps with the confusion.
Sorry, I meant nobody knows why they continue to work on it. It's pretty clear they're never going to get volume production out of it.
Which is why the rumor gained traction.
TSMC's "7nm" is actually a step behind Intel's 10nm but it wouldn't surprise me if Intel's 10nm which will be used will end up being a bit wider than initially expected - in the same way their 14++ is wider than 14+ and 14. That being said there's zero chance of Intel's process being in any way compatible with TSMC's.
Unlikely as they have to increase density to make such statement and it won't be possible on 14nm.
Nobody will have any advantage soon and everyone will struggle with newer processes. Transistor scaling is ending, we're almost at atomic levels now. As for Intel "succeeding the performance crown to AMD", well, let's see how AMD will fare on 7nm first - and if they'll be able to actually outperform Intel's 14nm CPUs. Then again Intel can always use the same TSMC foundry which AMD is using - AMD don't have a unique production line completely to itself anymore.
Honestly, if both companies could be fairly equal in performance when things completely stall out, it'd probably be best case scenario for everyone. Would result in best competition/prices.
AMD still has a lot of catching up to do for this. It's also rather naive to think that equal process would result in equal performance - ATI/AMD and NV has been using same processes for their GPUs since forever, didn't stop them from falling flat on their faces with new products from time to time. The market will still be competitive, it's just that with CPUs the end of scaling likely means the end of rising of number of CPU cores - but since we'll probably hit 16 or 32 on a consumer die by that moment this shouldn't be much of an issue.
It is factually true that Intel's process technology labels have remained "more honest" than those from other foundries, i.e. the sizes correlate much more closely with the actual dimensions. We hear this repeatedly from industry experts, hardware reviewers and so on. This has nothing to do with defense forces or fanboys. I'm not sure why you'd even be upset over that kind of thing since it's not an indictment of AMD (it's the foundries, not AMD or Nvidia that decide on the nomenclature).
They are both targeting the same ballpark. The precise perf of both nodes won't be known until we have comparable products on the market.
It was supposed to be better actually. But it's hard to say what will be its comparative state once it will be used for mass production now.
You're putting words into my mouth there. I never said equal processes = equal results. I simply said it'd be ideal if they could be close in performance when shrinks died out. Zen 2 obviously needs to make up a little more ground and AMD needs to find a way to match Intel's clock speeds, but if they can, it's a best case ending point. Contrast to where AMD was 5 years ago. That would have been an awful point (for consumers) to stall out at.
My point is that it never was about equal processes. The simple fact that all chips will eventually be made on a similar sized process doesn't mean that these chips will provide the same performance or even perf/dollar.
