Xbit Labs: AMD-> Improvements of Next-Generation Process Technologies Start to Wane.

[StrangerGuy]

expect huge price increase from Intel side.


I think this is another milestone in the death of x86.

They do that, get huge backlash, while rest of the masses continues to watch kitten videos on Youtube and say "why I would even want a new PC?".

 

[cbn]

They do that, get huge backlash, while rest of the masses continues to watch kitten videos on Youtube and say "why I would even want a new PC?".

Intel and AMD would have to target enthusiast and Server with some kind of higher performance CPU.

As it stands now, however, I don't think Intel could charge much more for Extreme Edition CPUs even if they wanted to. Hexcore? Why do even enthusiasts need that many cores?

If anything I'll bet Intel's Extreme edition market will shrink (if it isn't shrinking already) as people move to unlocked mainstream chips. <----LOL, I guess Intel could get rid of these if they wanted to and force people to step up to $1000 hexcore and octa-core E-type CPUs (just to get four fast cores). Maybe Intel will come up with some kind of justification too....."We need the unlocked multiplier chips to be EE or enthusiast socket only because the die layout on the new non-GPU chips is more conducive to cooling....etc....etc....etc...."

 

[zlejedi]

If they go this way soon Intel will have real 8 cores using die space of AMD pseudo 8 core Bulldozer and will be running at similar frequencies if they stop investing in shrinking nodes.

And well they can innovate - but what will stop intel from simply copying the idea and then outperforming them using superior tech ? Will they keep innovating every few years ?

 

[cbn]

Nvm.

 

[tweakboy]

Their all talk, I can talk too,,,,,, all talk and no show, well no kidding as tech gets better the fabrication will diminish to eventually 1nm by 2020 lolol gl,

 

[pelov]

Their all talk, I can talk too,,,,,, all talk and no show, well no kidding as tech gets better the fabrication will diminish to eventually 1nm by 2020 lolol gl,

How the hell do you have over three thousand posts here and someone hasn't banned you yet? Dear lord...

 

[Borealis7]

i think someone with around 30 posts is going to get banned...

aaaaanywho, their not saying anything we haven't heard before (from Intel or IBM). once Silicon is exhausted they'll have to switch to new materials and lithography. like i said earlier, Graphene is a likely candidate and relatively easy to produce (although maybe not mass-production). we could see a period where there is a very sudden and sharp rise in component cost due to transition to these new materials.
but that's all YEARS ahead. By that time, there's sure to be a cost-effective solution.

 

[Lonbjerg]

i think someone with around 30 posts is going to get banned...

aaaaanywho, their not saying anything we haven't heard before (from Intel or IBM). once Silicon is exhausted they'll have to switch to new materials and lithography. like i said earlier, Graphene is a likely candidate and relatively easy to produce (although maybe not mass-production). we could see a period where there is a very sudden and sharp rise in component cost due to transition to these new materials.
but that's all YEARS ahead. By that time, there's sure to be a cost-effective solution.

Intel has the last man years used ~$5 billion per year in R&D for "post-silicon".
My guess is that they are ready to ditch silicon ~2017.

 

[pm]

My guess is that they are ready to ditch silicon ~2017.

And switch to what?

 

[Lonbjerg]

And switch to what?

Ask Intel, the tidbits I have spotted seems to indicte they have a plan...they take their proces technology very serious.

 

[meloz]

I think this is another milestone in the death of x86.

All those predicting the death of x86 are going to look mighty foolish as Intel continues to improve performance/watt by approximately 15&#37; every year.

The reports of x86's death are greatly exaggerated.

ARM licensees should be worried, not Intel. When low power variants of Atom are fabricated on 22nm process, many current users of ARM will hop onto the x86 train.

With persistent demand for increasingly more powerful 'smartphone' chips, the market is moving towards Intel. Intel will catch it gracefully and without much effort around 2015 or so.

 

[meloz]

And don't worry about Intel price gouging either. The latest AMD processors might not live upto our great expectations and AMD marketing hype, but if Intel increases prices too much AMD's processors will again become competitive on a performance per dollar metric. OEMs for one will rush and buy all that GloFo can produce. It is a self-correcting mechanism.

So Intel has rather limited room to increase prices. Even if AMD stays a little behind Intel, as long as AMD exists we have nothing much to worry about.

 

[brybir]

Ask Intel, the tidbits I have spotted seems to indicte they have a plan...they take their proces technology very serious.

A company like Intel certainly has a plan, or more likely multiple plans, stretching five to ten years out. Intel has a pretty significant academic presence in many of the best universities in the world and they fund quite a bit of basic research at these institutions and in house. They certainly have their ear to the floor so to speak and are likely very aware of the future options.

 

[Maximilian]

How the hell do you have over three thousand posts here and someone hasn't banned you yet? Dear lord...

How the hell do you have over three posts here and someone hasn't banned you yet? Dear lord...

 

[pelov]

How the hell do you have over three posts here and someone hasn't banned you yet? Dear lord...

Do me a favor, go through his recent post history then go through mine.

 

[Idontcare]

Do me a favor, go through his recent post history then go through mine.

Actually, what he was referring to is that you ought to do yourself a favor and read through the posting guidelines, paying special attention to the following excerpts:

1) No trolling, flaming or personally attacking members. Deftly attacking ideas and backing up arguments with facts is acceptable and encouraged. Attacking other members personally and purposefully causing trouble with no motive other than to upset the crowd is not allowed.

We want to give all our members as much freedom as possible while maintaining an environment that encourages productive discussion. It is our desire to encourage our members to share their knowledge and experiences in order to benefit the rest of the community, while also providing a place for people to come and just hang out.

We also intend to encourage respect and responsibility among members in order to maintain order and civility. Our social forums will have a relaxed atmosphere, but other forums will be expected to remain on-topic and posts should be helpful, relevant and professional.

We ask for respect and common decency towards your fellow forum members.

Stupidity is not a crime, but being an ass is.

 

[exar333]

AMD is essentially saying here that performance and process do not matter that much. If this is true, then there are only two real players left, Intel and ARM. Intel for performance and process advantage, and ARM for affordability and power usage. AMD is left with no niche and no reason for someone to purchase them.

 

[soccerballtux]

How the hell do you have over three thousand posts here and someone hasn't banned you yet? Dear lord...

tweakboy is an upstanding citizen very important very interested in doing only what is very best for these forums and their members you would do well to pay him respect

 

[cbn]

once Silicon is exhausted they'll have to switch to new materials and lithography. like i said earlier, Graphene is a likely candidate and relatively easy to produce (although maybe not mass-production). we could see a period where there is a very sudden and sharp rise in component cost due to transition to these new materials.
but that's all YEARS ahead. By that time, there's sure to be a cost-effective solution.

Yes, it will be interesting to see how this material is deployed for computer chips. Unfortunately it sounds like it still has a good way to go.....

http://www.eetimes.com/electronics-news/4213661/Intel-s-Gargini-sees-tunnel-FET-as-transistor-option

When asked about graphene in the channel, another hot research topic, he said: "It's perfect across 100-microns but very imperfect on a 12-inch wafer; too many defects. It's going to take 15 years to get it right."

 

[cbn]

ARM licensees should be worried, not Intel. When low power variants of Atom are fabricated on 22nm process, many current users of ARM will hop onto the x86 train.

With persistent demand for increasingly more powerful 'smartphone' chips, the market is moving towards Intel. Intel will catch it gracefully and without much effort around 2015 or so.

Well Qualcomm currently has Intel beat on process tech for smartphones with the launch of 28nm Krait CPU Snapdragons in 1H 2012.

Maybe "Super Wifi" carriers will help Intel? Of course, this type of handset market would mostly likely be unsubsidized if it does end up existing....which brings up the question of how affordable Intel can make their Smartphone SOCs.

http://www.digitimes.com/news/a20111216PD209.html

Intel to struggle to see general adoption of Medfield in 2H12, say Taiwan makers

Monica Chen, Taipei, Joseph Tsai, DIGITIMES [Friday 16 December 2011]

Intel will launch 32nm Medfield specifically for use in smartphones and tablet PCs in the first half of 2012, but will struggle to win general adoption of the platform by vendors which have offered smartphones, according to Taiwan-based makers.

Because Intel did not establish close partnerships with first-tier smartphone vendors, while notebook vendors, which have smartphone product lines, are conservative about the new platform because of the negative experience with Intel's mobile Internet devices (MID), these factors are expected to pose strong difficulties for Intel to enter the smartphone market.

Intel has been working on improving its Atom SoC for smartphones and tablet PCs, and expects the new chip's power consumption will drop below 10W with related manufacturing process to also advance to 32nm Saltwell and then 22nm Silvermont, followed by 14nm Airmont within the next three years.

Although Samsung Electronics reportedly will launch a smartphone adopting Medfield and Android 4.0, and will sell it through Sprint, the sources pointed out that the cooperation between the two firms is because Samsung does not want to be limited by a single platform and therefore is more aggressive in developing products with different platforms, but the same situation may not apply for other first-tier smartphone vendors.

As far as 22nm FinFET goes, I thought that tech was mainly for high power?

 

[cbn]

And don't worry about Intel price gouging either. The latest AMD processors might not live upto our great expectations and AMD marketing hype, but if Intel increases prices too much AMD's processors will again become competitive on a performance per dollar metric. OEMs for one will rush and buy all that GloFo can produce. It is a self-correcting mechanism.

So Intel has rather limited room to increase prices. Even if AMD stays a little behind Intel, as long as AMD exists we have nothing much to worry about.

What you are saying makes a degree on sense to me back when the market was completely dominated by Big Windows and Intel vs AMD.

But now both those companies have ARM to deal with.

Intel may very well lower prices to compete with ARM more in the future. The resulting price compression could squeeze AMD out unless they are able to find some type of unique niche.

 

[IntelUser2000]

Well Qualcomm currently has Intel beat on process tech for smartphones with the launch of 28nm Krait CPU Snapdragons in 1H 2012.

You can't really say 28nm is an advantage. It's just a designation. It's a advancement from TSMC's own 32nm, but not directly comparable with Intel's 32nm.

Looking at details, TSMC may have some density advantage, and they might also have better tools for low power circuitry, though its to be seen if that'll remain true in the 32nm generation. Intel's transistors can drive much more current(45nm Intel is equal/better than TSMC 28nm).

http://www.digitimes.com/news/a20111216PD209.html

I find it funny how they take speculations and assume its real, then they take that and make a bigger story out of it.

Only thing we really know about Intel's plan for smartphones in 2012 is that they promised they will do it. Nothing about specific company collaborations(like in this case, Samsung).

 

[Lonbjerg]

Well Qualcomm currently has Intel beat on process tech for smartphones with the launch of 28nm Krait CPU Snapdragons in 1H 2012.

Maybe "Super Wifi" carriers will help Intel? Of course, this type of handset market would mostly likely be unsubsidized if it does end up existing....which brings up the question of how affordable Intel can make their Smartphone SOCs.

http://www.digitimes.com/news/a20111216PD209.html



As far as 22nm FinFET goes, I thought that tech was mainly for high power?

28nm is just a halfnode of 32 nm.
Intel dosn't bother with halfnodes.
Hence why they go 32nm -> 22nm.

Going 28nm dosn't mena you "beat" anyone on process technology...fo them to do that...they would have to have gone 22nm.

Which is waht Intel is doing...and going 3D trigate at the same time.
They (Intel) are at leat 1½ years ahead the competition.

I mentioned this before in this very thread (Nodes vs halfnodes)

 

[Idontcare]

As far as 22nm FinFET goes, I thought that tech was mainly for high power?

I think you may be thinking of "high cost" when you say "high power". Finfets are ideally suited for low(er) power devices, its scaling them up for high power that is the challenge.

But making them is not a low-cost adventure owing to the sizable upfront R&D expense required to develop them into a manufacturable device.

It is with this in mind that the comparisons are more readily drawn to the lower barrier to entry of developing low(er) planar cmos process tech at the continuing smaller nodes.

If you don't have much cash to throw at developing finfets for 22-20nm then sticking with planar CMOS is still viable provided you aren't directly competing with someone who has gone ahead and made their process all the more lower power with Finfets.

 

[cbn]

Going 28nm dosn't mena you "beat" anyone on process technology...fo them to do that...they would have to have gone 22nm.

Which is waht Intel is doing...and going 3D trigate at the same time.
They (Intel) are at leat 1½ years ahead the competition.

Yes, I stipulated process tech "for smartphones" because I knew Intel would also have 22nm Ivy Bridge out 1H 2012.

One thing that impresses me about Qualcomm is that they are able to justify the most expensive 28nm TSMC wafers before any other ARM smartphone chip designer. I am not exactly sure why this is? A pure guess would say Qualcomm's market position and overall demand allows this to happen?

In contrast, Intel appears to be delaying their cutting edge nodes for smartphone SOCs and other lower margin products (eg, Ivy Bridge celeron won't debut until 2013).

Maybe this will change in the future? But in order for this to happen I'd imagine Intel will have to build up demand for their Smartphone SOC. How this will happen I have no idea. It appears to me x86 is not a strength in this regard and therefore I have theorized non-CPU SOC techs would allow Intel to flank ARM. However, I have no clue what those non-x86 techs could be?

Maybe integration could be one of Intel's major selling points?

Take for example this iphone 4S logic board-->http://www.ifixit.com/Teardown/iPhone-4S-Teardown/6610/2 <----That has a lot of chips on there! I count fifteen, in total, front and back. I'd imagine being able to consolidate that mess into one chip (or perhaps stacked chips in 3D configuration) would allow for a larger battery in the smartphone chassis....this in turn could enhance the run time metric (all other things being equal).

It would be interesting to see what Intel could put together at 14nm FinFET with Airmont. But like I said, what will their strategy be? How will they get developers to build up apps when I read posts claiming the Android Dalvik Virtual machine is optimized for ARM/Neon? Native Development Kit for x86 would help with performance, but where will the handset volume come from?

The way things appear to me, this whole story appears to be a classic "chicken and the egg" one. Intel needs volume to get developers interested, but volume won't happen unless developers are onboard. I guess they could throw money at the problem and sign developers, but isn't too much of that counter-intuitive to the process of making money?