[WCCF] AMD Contracts TSMC To Produce Zen At 16nm Amidst Concerns Over 14nm

[desprado]

So much for 14nm.Now they to change sacrifice for 16nm.

http://wccftech.com/amd-contracts-tsmc-produce-zen-16nm-woes-14nm-process-troubles-globalfoundries/

 

[VirtualLarry]

I'm not ever sure why people still post WCCF articles...

 

[desprado]

I'm not ever sure why people still post WCCF articles...

Many other sites also posted.
If you do not like it than you can try other source by using google.

 

[gdansk]

How could GloFo have 14nm difficulties? They licensed the process from Samsung and buy the machinery from the same suppliers (NANO LRCX AMAT KLAC ASML).

Barron's said, in May,

Well-publicized delays in 14 nm equipment deliveries were blamed on issues related to fab-readiness timing and equipment installation delays, not yield problems.

Given the process is proven, this should be a very fast ramp-up after equipment is setup. GloFo reported volume production at Malta 8 in July. In total, I rank as "unlikely" for Zen to be a TSMC only chip.

 

[desprado]

How could GloFo have 14nm difficulties? They licensed the process from Samsung and buy the machinery from the same suppliers too. They even reported volume production in July.

They are lack of funding from Dubai.They also hold many share of AMD.

 

[hawtdawg]

How could GloFo have 14nm difficulties? They licensed the process from Samsung and buy the machinery from the same suppliers (NANO LRCX AMAT KLAC ASML).

Barron's said, in May, Given the process is proven, this should be a very fast ramp-up after equipment is setup. GloFo reported volume production at Malta 8 in July. In total, I rank as "unlikely" for Zen to be a TSMC only chip.

Samsung isn't shipping anything high-performance on 14nm.

 

[NTMBK]

If it means that they aren't subjected to GloFo's whims, then good. Not sure how this works with the WSA, though.

 

[ShintaiDK]

How could GloFo have 14nm difficulties? They licensed the process from Samsung and buy the machinery from the same suppliers (NANO LRCX AMAT KLAC ASML).

Barron's said, in May, Given the process is proven, this should be a very fast ramp-up after equipment is setup. GloFo reported volume production at Malta 8 in July. In total, I rank as "unlikely" for Zen to be a TSMC only chip.

Samsung couldnt even deliver the A9 yet. GloFo couldnt deliver anything and was completely dumped as supplier.

Also just because they license it is far from being the same. GloFo uses different tools.

Its quite clear TSMC is the best after Intel.

 

[AtenRa]

The WCCF article is taken from the EXPreview article that speculates of the swift to 16nm TSMC because of the latest ZEN delay for Q4 2016.

translated link to English

https://translate.google.com/translate?sl=zh-CN&tl=en&js=y&prev=_t&hl=en&ie=UTF-8&u=www.expreview.com%2F43087.html&edit-text=

The funny thing is that in July 2015, GloFo said they already have started producing chips on 14nm LPE in mass production.

http://www.kitguru.net/components/g...14nm-chips-for-our-customers-yields-on-track/

Now they cannot produce any more and AMD will go to TSMC for ZEN ??? Did anyone thought about AMD would have to redesign the ZEN SKUs for the TSMC 16nm FF++ process and make new masks and spend time and resources they dont have ???
So even if TSMC 16nm FF++ is ready a little earlier than GloFos 14nm LPP, say 6 mothns ??? the entire redesign, mask creation and evaluation for the new 16nm at TSMC will delay ZEN even further than stick to GloFo 14nm LPP.

Really people think before you type

 

[VirtualLarry]

The WCCF article is taken from the EXPreview article that speculates of the swift to 16nm TSMC because of the latest ZEN delay for Q4 2016.

Now they cannot produce any more and AMD will go to TSMC for ZEN ??? Did anyone thought about AMD would have to redesign the ZEN SKUs for the TSMC 16nm FF++ process and make new masks and spend time and resources they dont have ???
So even if TSMC 16nm FF++ is ready a little earlier than GloFos 14nm LPP, say 6 mothns ??? the entire redesign, mask creation and evaluation for the new 16nm at TSMC will delay ZEN even further than stick to GloFo 14nm LPP.

Really people think before you type

This. Rubbish article is rubbish. And you all got taken. Funny how the brigade is so quick to latch onto a negative AMD article.

 

[NTMBK]

The WCCF article is taken from the EXPreview article that speculates of the swift to 16nm TSMC because of the latest ZEN delay for Q4 2016.

translated link to English

https://translate.google.com/translate?sl=zh-CN&tl=en&js=y&prev=_t&hl=en&ie=UTF-8&u=www.expreview.com%2F43087.html&edit-text=

The funny thing is that in July 2015, GloFo said they already have started producing chips on 14nm LPE in mass production.

http://www.kitguru.net/components/g...14nm-chips-for-our-customers-yields-on-track/

Now they cannot produce any more and AMD will go to TSMC for ZEN ??? Did anyone thought about AMD would have to redesign the ZEN SKUs for the TSMC 16nm FF++ process and make new masks and spend time and resources they dont have ???
So even if TSMC 16nm FF++ is ready a little earlier than GloFos 14nm LPP, say 6 mothns ??? the entire redesign, mask creation and evaluation for the new 16nm at TSMC will delay ZEN even further than stick to GloFo 14nm LPP.

Really people think before you type

Is 14LPP suitable for a high performance CPU? Or would GloFo have to produce a special process variant just for AMD (again)?

 

[AtenRa]

Is 14LPP suitable for a high performance CPU? Or would GloFo have to produce a special process variant just for AMD (again)?

From the EXPreview article about GloFo 14nm LPE mass production and 14nm LPP tape out in late April 2015.

https://translate.google.com/translate?sl=zh-CN&tl=en&js=y&prev=_t&hl=en&ie=UTF-8&u=http%3A%2F%2Fwww.expreview.com%2F40242.html&edit-text=

According to official disclosure of the progress of their 14nm LPE process in January this year have been certified, and this month the yield level has reached the level of mass production. 14nm LPP process is also underway, it is expected to begin this year's Q3 quarterly certification, in 2016 Q1 quarter production. In addition, GF also provides a process for the 14nm LPP more features, including additional IP core, 2.5D and 3D TSV packaging, advanced DFM (design for manufacturing) and advanced EDA reference processes.

14nm LPP process has a variety of chip flow sheet, not only mobile processor, as well as some high-performance large core product, which is the more worthy of attention. AMD next-generation process of the CPU and GPU will choose 14nm FinFET process, GF is their old friend, these high-performance large-core processor AMD ultimately estimate of envy.

The 14nm LPP is more than enough for AMD high performance Chips. The process also have High Density Libraries if anyone wants them.

 

[ShintaiDK]

We know that was BS. GloFo lost its A9 allocation because they couldnt deliver. Funny enough Samsung is losing allocation fast as well because they cant deliver either.

As always GloFo is only good at slides and PR.

 

[frozentundra123456]

Well, if you read the article carefully, they make clear it is all speculation. All the statements are carefully nuanced with "allegedly". So who really knows.

The title of the thread should really reflect this though.

 

[sm625]

Well, if you read the article carefully, they make clear it is all speculation. All the statements are carefully nuanced with "allegedly". So who really knows.

The title of the thread should really reflect this though.

It does. That's what the [WCCF] tag means.

 

[Dresdenboy]

Well, if you read the article carefully, they make clear it is all speculation. All the statements are carefully nuanced with "allegedly". So who really knows.

The title of the thread should really reflect this though.

Don't try. Some brains in this forum have an integrated logic allowing any speculation through as fact and filtering facts proving otherwise as long as it fits the own view of the world.

There is research going on, but those affected need some Freudian Uber-me like capabilities to observe their own behaviour in this regard.

 

[Phynaz]

How could GloFo have 14nm difficulties? They licensed the process from Samsung and buy the machinery from the same suppliers (NANO LRCX AMAT KLAC ASML).

Barron's said, in May, Given the process is proven, this should be a very fast ramp-up after equipment is setup. GloFo reported volume production at Malta 8 in July. In total, I rank as "unlikely" for Zen to be a TSMC only chip.

I could give you my method of building a house, and you could buy the same tools as me. Doesn't mean you could build a house.

Designing a manufacturing process is half the battle - maybe even less. You then have to implement that process, and implementation is hard.

This could explain the delay of Zen, as switching to TSMC would require running through physical layout again.

 

[mrmt]

It will be nice for AMD execs to explain why they had to make an amendment in the WSA to remove the waiver that allowed them to manufacture CPUs at GLF but they didn't bother to make another amendment to reinstate, or get a new waiver. They'll be missing Rory's tap-dancing skills.

If they did that we'll be watching AMD BoD reaching new heights in terms of unprofessionalism and reckless management.

 

[Fjodor2001]

How could GloFo have 14nm difficulties? They licensed the process from Samsung and buy the machinery from the same suppliers (NANO LRCX AMAT KLAC ASML).

From the Wccf article:

As such GloFo is reportedly attempting to retool its 28nm equipment for 14nm production. This retooling however has been considerably slow and is the reason behind the recent delays in bringing volumes and yields up to snuff by the pure-play foundry.

I.e. GF is trying to reuse some of their old 28 nm equipment for 14 nm, where as Samsung likely is using a new setup for 14 nm.

 

[Arachnotronic]

From the Wccf article:

I.e. GF is trying to reuse some of their old 28 nm equipment for 14 nm, where as Samsung likely is using a new setup for 14 nm.

IDC told us this already

 

[Fjodor2001]

IDC told us this already

Worth repeating, since the issue was brought up again... Also, now we have a source for the claim...

 

[erunion]

Worth repeating, since the issue was brought up again... Also, now we have a source for the claim...

IDC is about a million times more credible than WCCF

 

[dark zero]

So GloFo is an absolute dissaster to the point to make AMD to rely on HiaTusSMC... Poor AMD and nVIDIA. They will.get late on the party just when Intel enters with Kabylake

 

[jpiniero]

GloFo is the worst. I can believe it although at the same time I doubt AMD has the money to convert to TSMC. GloFo is given a share of the iPhone contract on a silver platter and can't deliver?

Speaking of Apple and GloFo, GloFo being GloFo was probably the cause of AMD losing the Macbook contract during the Llano years.

 

[DrMrLordX]

The 14nm LPP is more than enough for AMD high performance Chips. The process also have High Density Libraries if anyone wants them.

Wait. 14nm LPP with HDL? You could . . . you could do some really amazing stuff with that kind of density.