(Semiwiki) Intel 14nm Delayed Again?

[NTMBK]

Semiaccurate has done a rather good writeup of this situation: http://semiaccurate.com/2014/02/19/sky-falling-intels-14nm-broadwell/

The bottom line is this:

So in light of all this, all we can say is, “deep breath people, chill”. There is a delay of Broadwell, it is process related, but it is exactly what Intel said it would be over a month ago. The process is delayed a quarter, Broadwell is still 2H/2014 even if that is a serious blurring of the lines, and nothing has changed. Anyone who thinks this is something new and shocking is simply ignorant of how things work in semiconductors.

If even Charlie is saying that Intel is telling the truth and it's going according to plan, then I think we can assume it'll be fine.

 

[Fjodor2001]

Hmmm.... is that really true? When Intel communicated the delay in October 2013, did they say that no Broadwell CPUs we're going to be released until 2014Q4, and that almost all models wouldn't be released until 2015?

 

[bullzz]

@Fjodor2001 - no, they said broadwel will be released in H2 14. now H2 cud mean Q4. but i think we will see products late Q4

 

[Fjodor2001]

So that way they could easily "hide" another 3-6 months delay, that they were not aware of back in October 2013? I.e. if they intended most products to be released Q3, that now won't be released until 2014Q4 (some) and 2015Q1 (most).

 

[witeken]

It would be great if Intel launched Broadwell at Computex. I think that was their original plan. I wonder what they're now going to do.

 

[OCGuy]

Wow people really post fake roadmaps from terrible Chinese rumor sites that contain spelling-error red underlines? :biggrin:

Anyway, this should most likely put to rest any doubt the Broadwell is coming to desktop, with the first parts showing up this year in one way or another.

http://www.bit-tech.net/news/hardware/2014/02/21/intel-broadwell-nuc/1

 

[NTMBK]

Wow people really post fake roadmaps from terrible Chinese rumor sites that contain spelling-error red underlines? :biggrin:

Anyway, this should most likely put to rest any doubt the Broadwell is coming to desktop, with the first parts showing up this year in one way or another.

http://www.bit-tech.net/news/hardware/2014/02/21/intel-broadwell-nuc/1

I don't think that Y- and U-series ultrabook/tablet processors are really what people had in mind when they said "desktop"! No >30W TDPs, no more than 2 cores, and soldered down.

But yes, Broadwell should be very nice for NUC-style PCs, as they can fit more performance into the same chassis. It'll be interesting to see if there's a Broadwell Gigabyte Brix Pro.

 

[witeken]

I just realized that every mainstream Broadwell Intel processor will have the same graphics capabilities as the Xbox One (if gen8 will have 40% higher performance with the same number of EUs, at 1200MHz). So a NUC with GT3 will be very interesting.

 

[cytg111]

I just realized that every mainstream Broadwell Intel processor will have the same graphics capabilities as the Xbox One (if gen8 will have 40% higher performance with the same number of EUs, at 1200MHz). So a NUC with GT3 will be very interesting.

If it is running against the same kind of gfx API as the Xbox One?
Then again, compared to something like mantle and whatever api xbox1/ps4 is using, the broadwell cpu may have the raw muscle power to drive regular directx to similar performance levels (or higher?)

 

[Homeles]

Parity with the XB1 seems a bit optimistic to me.

 

[witeken]

I mean parity on raw GPU power, ~1.3gflops, depending on performance increase from gen8 and clock speed. Optimized games for X1 will obviously still have an advantage, but I think it's a nice achievement for Intel to make the X1 GPU already almost obsolete against IGPs within 1 year.

If the DirectX equivalent of Mantle becomes widespread, the performance advantage of X1 will decrease even more, while the CPU is also superior.

 

[Nothingness]

I mean parity on raw GPU power, ~1.3gflops, depending on performance increase from gen8 and clock speed. Optimized games for X1 will obviously still have an advantage, but I think it's a nice achievement for Intel to make the X1 GPU already almost obsolete against IGPs within 1 year.

Do you really think raw FLOP/s is all there is to GPU? Should we talk about Larrabee? :biggrin:

PS - You meant TFLOPs not GFLOPs.

 

[Homeles]

Do you really think raw FLOP/s is all there is to GPU? Should we talk about Larrabee? :biggrin:

PS - You meant TFLOPs not GFLOPs.

That's definitely not what he was saying.

 

[Nothingness]

That's definitely not what he was saying.

You meant I misread this:

I just realized that every mainstream Broadwell Intel processor will have the same graphics capabilities as the Xbox One

Followed by this:

I mean parity on raw GPU power, ~1.3gflops, depending on performance increase from gen8 and clock speed.

Well for me it was at the very least ambiguous, hence my question.

 

[AtenRa]

Flops alone means nothing. You can have a GPU with half the flops of another GPU and still be faster in gaming.

 

[Hugo Drax]

Intel 14nm Delayed Again?


Hmm, no Broadwell in 2014 then ?? If that rumor is true, then 14nm ATOM may be pushed to 2015 as well.

Intel has a glut of inventory. There is no need to rush to broadwell.

 

[NTMBK]

Intel has a glut of inventory. There is no need to rush to broadwell.

In markets where Broadwell gives no great advantage (i.e. desktops), sure. But Broadwell gives thinner devices and longer battery lifes, helping x86 tablets and ultrathin laptops compete with ARM devices, and as such I'm sure Intel are getting it out as fast as they can.

 

[witeken]

When do you think Intel will introduce III-V materials?

If we extrapolate Intel's current progress, there will be a second generation Si/Ge or Ge (7nm), followed by III-V at 5nm. Intel once suggested that they might skip 10nm (won't be the case), so maybe they had the idea to skip SiGe or Ge and immediately go to III-V at 7nm? What do you think?

 

[mrmt]

Intel has a glut of inventory. There is no need to rush to broadwell.

Really, how so? Where's that reflected on Intel balance sheet or earning reports?

 

[Homeles]

Really, how so? Where's that reflected on Intel balance sheet or earning reports?

Intel is not AMD. They don't delay upcoming products because of excess inventory.

 

[mrmt]

Intel is not AMD. They don't delay upcoming products because of excess inventory.

AMD did that because they can't stop manufacturing chips due to the WSA quota, but Intel? If they had an inventory glut now they would simply reduce fab utilization and get inventory in line with their targeted levels. Intel having an inventory glut would not have been a question of missing sales forecasting, but also not adjusting production levels, and there's no way they can do that without reflecting in their balance sheet:

Sales forecasts would be lower if they were reducing fab utilization, inventory would balloon if they were having a glut and not adjusting production, but keeping inventory within the company, or margins would crash if they stuffed the channel with product. I don't see any of these elements in Intel financial information.

 

[SiliconWars]

Altera to switch 14nm chip orders back to TSMC.

http://www.digitimes.com/news/a20140304PB200.html

Meanwhile, TSMC has managed to ramp up the yield rate of its 20nm process to 50% recently, which will allow TSMC to enjoy brisk sales in the next three quarters.

Oh, and 450mm has been pushed out another 5 years - https://www.semiwiki.com/forum/content/3241-intel-450mm-delayed-until-2023.html

 

[Venomous]

Intel is finally going to slow down the train, I am assuming. Between SB/IB/HW being more than "good enough" for basically everything, causing even enthusiasts to question the need to upgrade, the increasing costs of developing each new process, as well as the global outlook for PCs, predictable "tick tock" is dead.

Fab 42 wasn't meant to be empty. Sony just ended VAIO as we know it. Global PC sales are on a 6-quarter decline streak. Today's headline from an "emerging" BRICs market:

PC sales slump 19%, manufacturers look for exit

http://timesofindia.indiatimes.com/...urers-look-for-exit/articleshow/30248511.cms?


I will personally keep an eye on the R&D expenditure. If we see that drop, then things could start to slow down even further in the next decade.

Just watch and see the 14nm SoC process with Atom be released in 2014 as stated. Not too hard to connect dots.

That was an interesting article. Once I saw the total sales of pc's vs tablets and smartphones, it makes SoC Atom dropping this year very plausible.

 

[Homeles]

Altera to switch 14nm chip orders back to TSMC.

http://www.digitimes.com/news/a20140304PB200.html

Seems a bit extreme. 14nm isn't that late. Altera's loss, really.

Oh, and 450mm has been pushed out another 5 years - https://www.semiwiki.com/forum/content/3241-intel-450mm-delayed-until-2023.html

Yeah, I saw that. Bad news for everyone, really. Especially consumers.

 

[bullzz]

intel has denied it. interesting
http://www.fool.com/investing/general/2014/03/05/did-intel-lose-altera.aspx