[Gulf News] Intel launching 10 nm parts in H1 2017

[Mondozei]

And do we believe them?

Broadwell was supposed to be launched in Q4 2013. It had a small and symbolic launch in december of 2014 but most Broadwell processors won't be out until this spring so Q2 2015.

I'm also amused how Intel's guy in the Middle East claims that they are still following Moore's law. Silicene, you can't come fast enough!

 

[oobydoobydoo]

When did people start using the words "pursuing moore's law"? Something tells me it didn't start until recently because I never recall hearing that phrase before ~2012 and even then, it didn't become really common until very recently as progress with new nodes became slower. It just strikes me as odd, you don't "pursue" laws. Laws are rules by which objects and systems operate.

2017 seems very, very early. Broadwell is still only available in theory, mostly.

 

[NTMBK]

When did people start using the words "pursuing moore's law"? Something tells me it didn't start until recently because I never recall hearing that phrase before ~2012 and even then, it didn't become really common until very recently as progress with new nodes became slower. It just strikes me as odd, you don't "pursue" laws. Laws are rules by which objects and systems operate.

Sure, but "Moore's Law" was never actually a scientific law. It was an empirical observation, which turned into a target/self fulfilling prophecy for the semiconductor industry.

 

[witeken]

People are hilarious.

*BDW was never supposed to launch in Q4'13. HVM was supposed to start then, but was delayed by 6-9 months because of yield problems.

*Intel started having 10nm start-up costs in Q4'14. No company would do that if they intended to have products out only in 2017 (let alone 2018 like people here assert).

 

[NTMBK]

*BDW was never supposed to launch in Q4'13. HVM was supposed to start then, but was delayed by 6-9 months because of yield problems.

Yes, and we have always been at war with Eurasia :\

 

[Exophase]

H1 2017 is plausible. H2 2016 I doubt. Process to process release has been slipping gradually for each of the past few nodes, and I feel pretty confident that we won't suddenly see a big reversal of that. I don't agree with the people who insist that the slips are a one time thing and consequently a less than two year gap will occur next time to make up for it. People were saying that when Ivy Bridge was delayed too (yes, it was delayed, remember?), and then could say that the transition to 22nm was a lot harder than the transition to 14nm would be because of FinFETs. But we all know how that turned out.

 

[Fjodor2001]

Intel has just retracted the statement here.

Intel retracts from its statement

Company now says the 10nm is under development

Dubai: Intel on Tuesday issued a statement to Gulf News that its 10-nanometre processors are under development, but for competitive reasons they are not discussing the timing.

The statement was a retraction from the company’s previous statement, which said that 10 nanometre processors were expected to be launched early 2017.

The new 10 nanometre reduces power consumption and boosts battery life.

I suspect the estimate provided by the Intel representative in the original article was correct, but maybe he said a little more in public than he was supposed to. 😱

 

[witeken]

Exactly what I said.

“We do think that we've been giving too much insight too far in advance and so we'll talk about 10nm some time in the next 12 or 18 months [from Nov '14] and when it's appropriate.” Stacy Smith, CFO Intel, IM’14

 

[Fjodor2001]

Exactly what I said.

It does not mean that Intel says the original release date estimate for 10 nm is still valid though, which some have claimed. It just means they will not comment on it in public (or they regret that they actually just did that some days ago 😉 ).

Otherwise we would be very close to release - Intel's statements from IDF 2013:

http://www.fudzilla.com/home/item/32458-intel-promises-10nm-2015

Intel just said that it can do 14nm in 2013
[...]
In 2015 it plans to ship 10nm chips
[...]
It goes beyond that, as in 2017 it plans to ship 7nm chips

Seems Intel has just been a tad too optimistic... :whiste:

 

[bullzz]

@Fjodor2001 - cant really beat GF and TSCM in optimism. how are those 14nm finfet chips doing 😀

 

[witeken]

@Fjodor2001 - cant really beat GF and TSCM in optimism. how are those 14nm finfet chips doing 😀

It's interesting that Intel has (mainly) issues with the triple patterning (small feature sizes) for yields, while TSMC's problems are related to the FinFET device Intel's been shipping with unprecedented yields and volumes since 2012 (the only difference between the 2 nodes is the FinFET, which takes half a year longer than the 1 year they expected).