intel delays 14nm?-process problems

[mrmt]

Originally Posted by piesquared

Besides stockholders to answer to, there is also the ARM stampede to battle. Using that tired old meme of intel not have any competition to explain intelss execution problemss has gotten ridiculous.

Since 2006 Intel has delivered four architectures and three process nodes, more than anyone in the industry, with only minor delays. More important, they are ahead of the curve, which means that whatever problems they have they don't have anywhere to look but themselves to find a solution.

That's a stark contrast with Nvidia Fermi woes, or AMD/GFL 32nm snafu, or IBM late introduction of their 32nm node, or Sun's Rock architecture. But some people seem to have one standard when judging Intel execution and another to judge the rest of the market execution.

[Hulk]

Originally Posted by mrmt

Since 2006 Intel has delivered four architectures and three process nodes, more than anyone in the industry, with only minor delays. More important, they are ahead of the curve, which means that whatever problems they have they don't have anywhere to look but themselves to find a solution.

That's a stark contrast with Nvidia Fermi woes, or AMD/GFL 32nm snafu, or IBM late introduction of their 32nm node, or Sun's Rock architecture. But some people seem to have one standard when judging Intel execution and another to judge the rest of the market execution.

Well stated. Intel has been leading the microprocessor industry (with a few minor setbacks) for decades. My respect for them grew tremendously when they were briefly overtaken by AMD in 2005 but came roaring back in 2006 with C2D never looked back. The point is that they have dealt with adversity before and passed the test. I wouldn't worry about Intel. I only hope AMD can stay relevant and that Apple will eventually pose a threat to Intel on the CPU side of things. As usual, we all benefit from healthy competition.

[soccerballtux]

Originally Posted by mrmt

Since 2006 Intel has delivered four architectures and three process nodes, more than anyone in the industry, with only minor delays. More important, they are ahead of the curve, which means that whatever problems they have they don't have anywhere to look but themselves to find a solution.

That's a stark contrast with Nvidia Fermi woes, or AMD/GFL 32nm snafu, or IBM late introduction of their 32nm node, or Sun's Rock architecture. But some people seem to have one standard when judging Intel execution and another to judge the rest of the market execution.

it's the big name.
They have the brightest minds, because they can pay for them.
If the brightest can't figure it out, we're doomed.

That said, I think this is some random guy starting a rumor out of nothing.

[ShintaiDK]

Originally Posted by StrangerGuy

Welcome to peak silicon!

Bohr says its secure basicly until 5nm. So abit of a way still.

[mikk]

At least two other 14nm fabs are in the works.

[Phynaz]

Nothing says process problems. But I suspect the op knew that.

[BD231]

Originally Posted by Lonbjerg

Google: stockholders

The value of your post: Zero

Speaking of useless posts ....

Intel had a die shrink this year d-bags, of course they're not going to have one next year.

[ShintaiDK]

Originally Posted by BD231

Speaking of useless posts ....

Intel had a die shrink this year d-bags, of course they're not going to have one next year.

Well factories will run production in late 2013 to the product launch in Q1 2014 for Broadwell and Airmont.

[Nemesis 1]

Originally Posted by piesquared

Not working? This could be the great equalizer.

http://www.xtremesystems.org/forums/...4nm-in-Ireland

Uh oh.

Mind pointing out were either article says anything like what your saying , You can't because it doesn't say any such thing . They are talking Intel ireland

[GreenChile]

Originally Posted by ShintaiDK

Well factories will run production in late 2013 to the product launch in Q1 2014 for Broadwell and Airmont.

Historically D1D is the only fab that runs production that early in the development cycle. HVM fabs start coming online many months later.

Pushing out the startup of an HVM fab usually would indicate Intel is forecasting a slower volume of 14nm products is needed. It does not necessarily indicate a delay in 14nm development. Although personally, I won't be surprised to see it delayed somewhat.

[2is]

Originally Posted by piesquared

This could be the great equalizer.

It's not.

[OCGuy]

This would be the OP's dream, but I highly doubt it.

The fact that we are even talking about this node for Intel, and its main rival can't even top a 45nm chip from 4 years ago is telling.

[Haserath]

Originally Posted by OCGuy

This would be the OP's dream, but I highly doubt it.

The fact that we are even talking about this node for Intel, and its main rival can't even top a 45nm chip from 4 years ago is telling.

Piledriver is overall better than Nehalem. The problem is Intel has increased efficiency a lot since Nehalem.

[Vesku]

Seems a response to market conditions, slowing market so slower ramp up.

[CHADBOGA]

Originally Posted by Haserath

Piledriver is overall better than Nehalem.

Which Nehalem?

Intel had the i7 970 Hex core selling cheap at the end.

[Ferzerp]

Originally Posted by Haserath

The problem is Intel has increased efficiency a lot since Nehalem

How is improving a problem? Do you want us to be stuck on 2008 tech, because that's what Nehalem is. 4 (or 4.5) year old tech.

[Eureka]

Originally Posted by Ferzerp

How is improving a problem? Do you want us to be stuck on 2008 tech, because that's what Nehalem is. 4 (or 4.5) year old tech.

Problem for AMD.

[iCyborg]

Originally Posted by CHADBOGA

Which Nehalem?

Intel had the i7 970 Hex core selling cheap at the end.

There were no hex-core Nehalems. i7 970 is a Gulftown.

[GreenChile]

Originally Posted by iCyborg

There were no hex-core Nehalems. i7 970 is a Gulftown.

Not exactly true. Nehalem is the code name for the microarchitecture. It does not reference any particular chip. It's true the 45nm variety of Nehalem microarchitecture did not have any hex core chips but the 32nm variety (also known as Westmere shrink) did. The i7 970 is still considered Nehalem microarchitecture.

[pablo87]

Intel has been spending capex like there's no tomorrow, betting on the come from where???

PC sales are losing vs tablet and smartphones, even the most enthusiastically reviewed "PC" on Amazon nowadays is the ARM based Chromebook...

Just to maintain volume status quo, intel will need AMD to go out of business...that's what I posted here a week ago, Barclay's said the same thing today (basically).

Which is why on the Q3 con call, Intel's CFO stated capex would be trimmed...in that context, delaying one or more of the fabs makes sense. Which begs the question, is TSMC's business model catching up to Intel's? It sure seems that way.

The irony: GF not doing well for many reasons but mainly because AMD isn't doing well, makes TSMC a stronger competitor for Intel.

[Hulk]

Originally Posted by ShintaiDK

Bohr says its secure basicly until 5nm. So abit of a way still.

Bohr also used a purely Newtonian physics model to predict the energy states of the hydrogen atom. Yeah it was right for hydrogen but that was about it. I'll stick with Schrodinger's wave equation.

Just kidding. I know what you mean.

[happysmiles]

Intel doesn't need AMD to go out of business, that doesn't stop the tablet/phone sales.

Adapt or die it is essentially, again Ivy Bridge and Haswell are majorly focusing on power consumption.

[iCyborg]

Originally Posted by GreenChile

Not exactly true. Nehalem is the code name for the microarchitecture. It does not reference any particular chip. It's true the 45nm variety of Nehalem microarchitecture did not have any hex core chips but the 32nm variety (also known as Westmere shrink) did. The i7 970 is still considered Nehalem microarchitecture.

Most people wouldn't call 3770K a Sandy Bridge, even though the same argument would apply. Anyway, if you go up a few posts, you'll see that the original comparison was to a "45nm from 4 years ago"...

[Borealis7]

Originally Posted by mikk

i have found the source of the 14nm problems. look at the picture: Intel has replaced "Israel" with "Ireland" and thus losing 4 FAB points. (28 vs 24)

but really, for Intel - the only problem with 14nm is the lack of incentives (read: competition) to make it work. i doubt it even makes economical sense to build an entirely new fab for this process plus converting 2 other fabs to 14nm, think about it as diminishing returns.
but they will do it anyways.

[Edrick]

Yay for AMD fanboys.