Filed under: 10nm, CPUs, Featured, Hardware, Intel, News, Sticky

Intel: 10nm Yield Is Ahead Of Expectation, 7nm Ponte Vecchio…

Intel has had a beautiful quarter and the firm is literally selling each chip it might perchance most likely most likely well presumably secure. Seems to be there had been just a few extra nuggets of files buried within the slides as correctly. Intel has officially stated that their 10nm yields are earlier than expectations and that they are planning as a minimum 9 10nm products for commence in 2020. The 7nm lead product, Intel Xe Ponte Vecchio GPU remains on-music for Q4’21.

Intel planning 9 10nm products in 2020, yields earlier than expectations

Intel is incredibly expert at bettering architecture while staying on the identical node, but inquisitive about 14nm has been spherical for a undoubtedly long time, the shift to 10nm is eagerly anticipated and is generally a turning level within the firm’s history. Whereas there had been some rumors earlier on (that Intel readily debunked) it is far no longer certain that 10nm HVM is anticipated to secure landfall in 2020. Intel has extra revealed that no longer ultimate are yields above expectation but they’ll be rolling out a total of 9 10nm products in 2020.

Listed below are the key takeaways:

  • 10nm yields are earlier than expectation.
  • Nine 10nm products are anticipated to commence in 2020.
  • Intel will lengthen its wafer skill by 25% in 2020.
  • 7nm Ponte Vecchio GPU no longer off beam for Q4 2021.

It might perchance well presumably furthermore appear that work on Xe-basically based Ponte Vecchio is going big – which is factual news for avid gamers on story of this tech would in the end trickle down within the gaming space as correctly. Intel’s Ponte Vecchio GPU for 7nm would be the necessary necessary milestone within the firm’s ambition to vary steady into a GPU participant and would constitute the necessary necessary diversification. Intel’s DG1 GPU is likely one amongst the 10nm products that will likely be launched later within the Twelve months. As for the diverse eight products, these are likely going to embody a mix of the CPU side as correctly as FPGA/AI basically based products.

Recap: Intel CEO reaffirms commitment to 7nm and explains what went defective with 10nm

Bob Swan has previously answered a ask about how Intel got steady into a local the place it has lost an enormous chunk of its CPU market fragment to AMD and is in a local the place it is far unable to satisfy quiz (here is in distinction to its passe philosophy of prioritizing err-ing on the side of caution and constantly having spare fab skill):

How we came is largely extra or less threefold, one we got lots faster than we anticipated and the quiz for CPUs and servers grew noteworthy faster than we anticipated in 2018. You’ll take into accout we came into 2018 projecting a ten% development and we grew by 21% development so the factual news effort is that quiz for our products in our transformation to a knowledge-centric firm changed into noteworthy increased than we anticipated.

Secondly, we took on a 100% market fragment for smartphone modem and we made up our minds that we would fabricate it in our fabs, so we took on noteworthy extra quiz.

And 1/three, to exacerbate that, we slipped on bringing our 10nm to lifestyles and when that happens you fabricate extra and additional performance into your final generation for us – 14nm – which plan there is a increased core depend and increased die dimension.

So these three – increasing noteworthy faster than we conception, bringing modems inner and delaying 10nm resulted in a local the place we didn’t have versatile skill. – Intel CEO Bob Swan

Whereas most of here is passe news, here is the necessary time that Intel has given a stable motive for why it is far no longer ready to satisfy skill – namely that it made up our minds to scheme smartphone modems in-home which in return meant that they weren’t ready to level of curiosity on the CPU side of things. It is furthermore a fairly believable explanation for why Intel can no longer even meet the quiz for 14nm anymore and as a alternative has to resort to extending 22nm products.

When asked particularly to express what went defective, Bob Swan candidly replied with an admission that Intel had gotten overconfident in its ability to beat the alternate licensed and suffered the consequence. Calling it “Scar tissue” here is the reason Bob gave:

The scar tissue undoubtedly started with Moore’s Law. Two times scaling part each two years and that’s the reason extra or less the easy rule of thumb. That is worked for a undoubtedly long time. And the transition from 22 to forty nanometers after which 14 to 10 we made up our minds that despite the reality the physics changed into getting extra fascinating we made up our minds to characteristic a increased bar for ourselves when it comes to performance. So the 22 to 14 is no longer a 2 times density, it changed into 2.Four and it changed into bumpy along the plan but it absolutely worked and that working gave us the self belief that for 14 to 10 why don’t we get rid of the scaling part as much as 2.7 if you appreciate that the implications of trying to procure extra and additional density and additional and additional performance [ you start to see the problem]

Secondly, we’re no longer going to examine to appreciate 2.Four scaling or 2.7 scaling as we deem about 7 nm, you know, we set aside 2.0 again in conserving with historical traits as we deem about 5nm, which would be our competitors three nanometer…

In our first 7nm product within the fourth quarter of 2021… I’d furthermore argue that another particular of just a few of the challenges spherical 10 nm is that you further or less realized be taught how to secure 14 greater along the plan and likewise you know, we have 14+,  14++ and and despite the reality that you have got got been on the identical node for four years now performance of the chips that continue to toughen as we trip to 10. – Intel CEO Bob Swan

With Intel chasing a 2x scaling part for 7nm and transferring to EUV as correctly, it looks admire the firm is all characteristic to introduce its first 7nm products (equal to TSMC 5nm) within the fourth quarter of 2021. Bob furthermore extra stated that he expects to hit 5nm (equal to TSMC 3nm) by 2H 2024.

The post Intel: 10nm Yield Is Ahead Of Expectation, 7nm Ponte Vecchio GPU On Music by Usman Pirzada regarded first on Wccftech.