Intel says Ivy Bridge delays are due to process, not lack of demand

[IntelUser2000]

Eh, I'll believe it when I hear it from a different source?

I doubt there will be issues in the long term, but I'm talking about Ivy Bridge. I don't think they can afford to hold back on chips when Tablets are really becoming relevant this year and they need to keep Apple as a customer. If low demand is true, they should do a firesale on Sandy Bridge to clear it out. But that's not what they are doing.

At CES they claimed 70 Ultrabooks are coming this year, while we only saw half a dozen of them. I wouldn't be surprised if 2/3rds of the designs are being built around Ivy Bridge, a second generation Ultrabook platform which will bring form factors and features(NFC, convertibles, touch screens) to really differentiate from Macbook Air, and be more than a thin laptop.

Just want to point to bulldozer... great overclocks on ES, highest overclock ever... bad yields, bad clocks, ...

Exactly. The theoretical top frequency might be much higher, but that's irrelevant for the vast majority of us.

 

[IntelUser2000]

I might have to revise this:

Xeon(E3?)/Desktop/Mobile Quad: Moved from April 8 to April 29

The articles didn't state desktop and mobile chips per se, I just assumed that from the wording. The date reference is to a Xeon, and only Xeons coming in that timeframe are Xeon E3's, which are based on LGA1155 chips. So it made sense to add desktop and mobile.

We have 3 possible dates on desktop and mobile quad core release: April 2, April 8, and April 29.

 

[denev2004]

OOOOOOOOps. Thats bad.
I scheduled to buy a notebook with IVY mobile in June....Sounds like I can only bought it in July.

 

[Nemesis 1]

I believe this is marketing only. AMD has only one chip that interest intel . and thats trinity. I do believe Intel wants AMD trinity to launch first. I know if I was Intel I wouldn't show IVB ahead of trinity. Let trinity come out than very shortly thereafter launch IVB.

 

[podspi]

I believe this is marketing only. AMD has only one chip that interest intel . and thats trinity. I do believe Intel wants AMD trinity to launch first. I know if I was Intel I wouldn't show IVB ahead of trinity. Let trinity come out than very shortly thereafter launch IVB.

What is your reasoning behind this? I find it hard to believe specifications aren't more or less fixed at this point...

 

[Idontcare]

I believe this is marketing only. AMD has only one chip that interest intel . and thats trinity. I do believe Intel wants AMD trinity to launch first. I know if I was Intel I wouldn't show IVB ahead of trinity. Let trinity come out than very shortly thereafter launch IVB.

Intel worries about what AMD is doing like AMD worries about what Via is doing. AMD has made itself irrelevant to the big picture thanks to spinning off their fabs and bungling the development of bulldozer.

4 yrs from now the conversation in this forum about AMD will be about as lively as the discussions on Matrox in VC&G.

 

[sm625]

Im surprised intel doesnt just wait till jan 2013. Milk that cow. Milk it.

 

[Idontcare]

Im surprised intel doesnt just wait till jan 2013. Milk that cow. Milk it.

Realize that the sole motivation for developing and releasing new nodes is to milk the cow.

If margins weren't better on a new node then the new node is intentionally delayed until the costs and margins are improved to the point where they are superior to the node it is replacing.

Intel isn't going to delay 22nm simply to "milk" the opportunity to sell lower margin 32nm skus. But they will delay it if they built more 32nm parts in Q4/11 than sold in Q1/12 and their partners are asking for extra time to draw down the inventory.

Nobody likes taking a writedown on inventory. There's no mystery in that.

However, supposedly Intel execs have expressly stated this delay is due to process tech maturity, i.e. yields. And one thing that has been conspicuously missing from Intel's usual bevy of "we are teh best on earth" marketing materials is their defect-reduction versus time graph showing off how well they've mastered 22nm in the fab.

In the run-up to each node, Intel likes to throw the following out to the investors:

Source

^ haven't seen Intel doing any of their usual marketing stuff on showcasing 22nm going swimmingly well. Makes me think this is not about inventory management and really is about process tech yield issues.

 

[Edrick]

I do not think it is a big mystery that Intel is having a tougher time with the 22nm node than they had with the last few previous nodes. And since there is no reason to rush out 22nm CPUs that may have some bugs, thy will take their time and do it right.

If AMD did not botch Bulldozer, and Intel was forced to get Ivy out fast, you can bet the house that Intel would have released them in Jan. 2012 (with a revision coming later).

 

[railven]

Is it possible this ties into that Article about Intel possibly opening up the fabs to outsiders?

If the 22nm might have been problematic, what better way then open up the next node to guinea pigs, sorry I mean customers, to work out the kinks?

 

[Edrick]

If the 22nm might have been problematic, what better way then open up the next node to guinea pigs, sorry I mean customers, to work out the kinks?

Yea, customers will pay Intel to be Intel's node beta testers.

 

[Rvenger]

4 yrs from now the conversation in this forum about AMD will be about as lively as the discussions on Matrox in VC&G.

I hope you are wrong IDC!

 

[Nemesis 1]

What is your reasoning behind this? I find it hard to believe specifications aren't more or less fixed at this point...

ya specs do seem to be inplace . intel has other opions . THe 77 watt can easily be pushed up to 95watts. At 32nm you could O/C the hell of of Igpu SB. 22nm should easily O/C to 2 ghz. It wouldn't surprize me in the least to see intel bring a higher performering IGP to market in stealth mode. I believe Intel wants to see trinity and than release. Intel can afford to hold back AMD can't . Llano was ok but didn't sell nearly as well as hardware sites and forum members talked up . Not even close to reality.

 

[Nemesis 1]

Yea, customers will pay Intel to be Intel's node beta testers.

Ya something like PH I and BD right?

 

[Nemesis 1]

Intel worries about what AMD is doing like AMD worries about what Via is doing. AMD has made itself irrelevant to the big picture thanks to spinning off their fabs and bungling the development of bulldozer.

4 yrs from now the conversation in this forum about AMD will be about as lively as the discussions on Matrox in VC&G.

Intel is looking in the rear view mirror. AMD can't hold trinity back they have to release to live to fight another day. Besides Both AMD and Imagination tech will be bought out this year. Q will buy AMD out . Intel will by VR tech. They all ready own 25%

 

[Nemesis 1]

I doubt there will be issues in the long term, but I'm talking about Ivy Bridge. I don't think they can afford to hold back on chips when Tablets are really becoming relevant this year and they need to keep Apple as a customer. If low demand is true, they should do a firesale on Sandy Bridge to clear it out. But that's not what they are doing

I have not read one thing that says apple will not get their chips from intel on time.

 

[IntelUser2000]

Is it possible this ties into that Article about Intel possibly opening up the fabs to outsiders?

If the 22nm might have been problematic, what better way then open up the next node to guinea pigs, sorry I mean customers, to work out the kinks?

It might not be much of a problem for those customers as they make FPGAs. TSMC does the same thing with their FPGA customers.

 

[Nemesis 1]

^ haven't seen Intel doing any of their usual marketing stuff on showcasing 22nm going swimmingly well. Makes me think this is not about inventory management and really is about process tech yield issues

So when was that chart released ? How long before SB release was that graph shown?

 

[Phynaz]

We ll just let them stockpile and slow down production, overall we make more money that way.

Which is a very expensive way to conduct business and would cause a manufacturing company to make less money, not more.

The holy grail of manufacturing is fully utilized lines, not factories sitting idle.

 

[Olikan]

heh!

first the HD shortage, now yields issues... is that hard to admit that desktop cpu is dead?

actually, desktop in general is dyeing

 

[God Mode]

heh!

first the HD shortage, now yields issues... is that hard to admit that desktop cpu is dead?

actually, desktop in general is dyeing

Saturated but not dead. Sooner or later, there will be a wave of people upgrading their S775/S1156 systems. I think a lot of people with SB systems are better off waiting for Haswell/Broadwell.

 

[Fjodor2001]

www.ft.com/cms/s/2/c3c3c2b6-602f-11e1-8de4-00144feabdc0.html
http://eetimes.com/electronics-news/4236944/Intel-s-Ivy-Bridge-delayed-says-senior-executive

Xeon(E3?)/Desktop/Mobile Quad: Moved from April 8 to April 29
Dual core i5/i7 mobile: Moved from May 13 to June 3
Core i3: Unchanged at June 24

Actually, the second article only says:

"Jim McGregor of In-Stat told EE Times that, according to his industry sources in Taiwan, Intel's Ivy Bridge server parts were only delayed from April 8 until April 29, though the dual core i5 and i7 parts for notebooks had been pushed out from a planned May 13th launch to June 3.

Core i3 parts would launch as planned on June 24, said McGregor."

It does not mention the Core i5/i7 Desktop Ivy Bridge CPUs being delayed. So will they still launch on time, i.e. April 8?

 

[Idontcare]

So when was that chart released ? How long before SB release was that graph shown?

Nearly 2yrs. That is what is disconcerting about the prospects of 22nm health. When they made their similar press release about 22nm it was notably lacking any indication of defect density management.

 

[meloz]

I doubt there are any technical issues with dual-core IVB. I suspect Intel is leaking 'reports' of delay to prevent a collapse of Sandy Bridge sales in the month(s) leading up to Ivy Bridge debut. The increased prices of HDDs have clearly reduced the global sales this quarter, particularly at lower end where dual core processors dominate and consumers are more sensitive to prices.

The 'issues' with dual core IVB could be a ruse by Intel to give themselves some time to clear out unsold processors (and chipsets, let's not forget).

 

[bradley]

Maybe Intel is slowly attempting to pull back their roadmaps and revise pricing tiers to reflect a lack of competition. I was initially surprised at how relatively cheaply the i5-2500K was priced for what it offers.