Intel Broadwell Thread

[witeken]

Care to elaborate...?

Sure..

Thanks very much. With the current schedule for the Broadwell ramp on 14 nanometers, what are your expectations for 10-nanometer timing? And will there be a follow-on 14-nanometer family after Broadwell or would you go straight to 10 nanometers?

Brian M. Krzanich - CEO
So let me try to answer your question. There’s a series of products that are on 14 nanometers. We always remember to do a Tick-Tock kind of products, plus we do a refresh of the Atom line typically, so we have Cherry Trail coming in at the end of this year and ramping in the first half. We have Broxton which is the follow-on Atom product as well and then the [Talk] [ph] or the real architectural shift on 14 nanometers is SkyLake which is also scheduled for '15. We have done no changes or shift to our 10-nanometer schedule but we won’t really talk about 10-nanometer schedules until next year.

 

[Fjodor2001]

Sure..

But the Browadwell tick is being skipped for most desktop SKUs....

 

[crashtech]

Maybe they figure "some" instead of "most" doesn't break their own rule.

 

[jpiniero]

But what's the point for the customer?

Lower power consumption. Core is already a super overkill for the vast majority of users, so going for lower power makes a ton of sense.... even for the desktop.

 

[swilli89]

Does anyone even really know if we are getting a Broadwell 4770k equivalent for desktop? This uncertainty is pretty odd to me I've gotta say.

 

[ame132]

Does anyone even really know if we are getting a Broadwell 4770k equivalent for desktop? This uncertainty is pretty odd to me I've gotta say.

yes
http://www.anandtech.com/show/7875/new-unlocked-iris-pro-cpu-broadwell

 

[crashtech]

There's been some FUD spread around that Broadwell-K won't actually be compatible with 9-series LGA 1150 boards. but when I looked for the original English source of the FUD at VR-Zone, it seems that something may have been lost in translation..? Information on this seems to be scarce. Has there been any official Intel confirmation as to in which boards socketed Broadwell will run?

 

[mikk]

Series 9 Boards should support BDW-K.

http://www.bit-tech.net/hardware/motherboards/2014/05/11/the-intel-9-series-chipsets-examined/1

Confirmed from Intel.

 

[IntelUser2000]

No that's a Chromebook, not a Notebook. Also, it's EOL since Broadwell will be released soon.

It's a Notebook without the cost of Windows. Since you are talking about the cost of the hardware, it matters.

But what's the point for the customer?

If the low-end Core and high-end Atom performs the same, its up entirely to Intel to supply the chips it wants to manufacturers.

See, certain decisions made by big super profitable companies like Intel don't always make sense, until you realize one thing - they do this to make much money as possible. I guess you can argue about the ethics of jacking up the price unnecessarily because they have practically no competition in that segment, but then again that's how they operate, and that's true with all companies.

So tick-tock is dead on desktop?

No, they are being flexible. I did say on Anandtech forums before that they skip certain delayed or lesser required segments so their process technology is up-to-date. If they have to wait for ALL segments to bring Broadwell, 10nm would be unnecessarily delayed. And they can't even consider reducing the lifespan of Broadwell to anything less than 12 months or else there would be no point of the work at all(manufacturers would skip it for the next one).

 

[Abwx]

There's been some FUD spread around that Broadwell-K won't actually be compatible with 9-series LGA 1150 boards. but when I looked for the original English source of the FUD at VR-Zone, it seems that something may have been lost in translation..? Information on this seems to be scarce. Has there been any official Intel confirmation as to in which boards socketed Broadwell will run?

Why would it be FUD.?.

Dare to give some proves first because if you have none its just plain deffamation, i posted a link that say so :

En effet Intel explique dans le diagramme ci-dessus que pour fonctionner, ces Broadwell nécessiteront une nouvelle tension dénommée VCCST (pour VCC Sustain).

I translate :

"Actualy Intel explain in the diagramm above that to work thoses Broadwell will necessitate a new voltage named VCCST (for VCC Sustain)."

It s explained in the article that only Haswell BGA have this voltage supply implemented but not the LGA versions while it will be mandatory on Broadwell LGAs.

http://www.hardware.fr/news/13310/broadwell-lga-incompatible-avec-lga-1150-actuelles.html

So actualy it s real infos that you are trying to denigrate using BS, so much for your pathetic FUD brandings, that s all you have as "argument".?.

 

[crashtech]

My, my, don't you take things a bit personally? I'm not arguing with you. I'd like to hear from someone else who does not share the same ethical concerns that you have about Intel.

Besides, Fear, Uncertainty and Doubt do not necessarily mean that I believe your link contains false information... so calm down.

 

[Abwx]

My, my, don't you take things a bit personally? I'm not arguing with you. I'd like to hear from someone else who does not share the same ethical concerns that you have about Intel.

Besides, Fear, Uncertainty and Doubt do not necessarily mean that I believe your link contains false information... so calm down.

You, should calm down, you gave a link to my post saying that it was FUD, wether or not your post was adressed to someone else doesnt matter since you made the accusation that I did spread false information, besides it s not my information but Hardware.fr s one, crazy that the fact that i m just posting what they are saying is assumed as being a fabrication from mine or ethicaly motivated...

Hfr article is dated 26/08/2013 , that s a little more than one month old, and the articles is not an interrogation, they says expressely in their title that :

Broadwell LGA incompatible avec les LGA 1150 actuelles

Les processeurs Broadwell au format LGA utiliseront toujours un Socket LGA 1150, mais ils seront par contre tout simplement incompatibles avec les cartes mères Socket LGA 1150 actuelles !

"Broadwell LGA incompatible with current LGA1150.

Broadwell processor in LGA format will indeed use a LGA1150 but they will simply be incompatible with current S1150."

Now you can always say that Hardware.fr is spreading FUD...

 

[TreVader]

I was really hoping to get he broadwell MacBook Air but it's looking like that actually might be a downgrade unless I'm misunderstanding broadwell turbo.

The i7 in the MacBook Air is a 4650U

http://ark.intel.com/m/products/75114/Intel-Core-i7-4650U-Processor-4M-Cache-up-to-3_30-GHz

How is Core M with the same amount of cache, much lower base and much lower turbo going to keep up with Haswell?

I really hope Intel has kept a secret SKU for apple that will help core M compete with haswell (currently clocked 600MHz higher than 5Y70). I want Broadwell to be good and I like haswell too, but I think apple will have something special for the MacBook line this year.

 

[jpiniero]

It looks like Apple is going to use Broadwell U instead and just wait until it gets released in early 2015.

 

[TreVader]

It looks like Apple is going to use Broadwell U instead and just wait until it gets released in early 2015.

This really shows you the state of intels delays. They are REAL and obviously a threat to Intel on OSX.


I'm very confused about people who talk about Intel having a current process lead. From what I can see, 22nm has far worse performance than 28nm in SOC form (bay trail fiasco) and only marginally better than 32nm sandy bridge in socket form.

Right now TSMC has 20nm products on the shelves. If Intel was capable of meeting their deadlines they would have a big paycheck from apple, but from what I can see Intel could learn a thing or two from TSMC about delivering a product on time.



Where is Broadwell? Im all for Intel in MacBook if they offer an advantage. Right now they are making apple miss deadlines. Broadwell looked so promising but I have a feeling Intel will do the usual butcher job and give 79 different skus with almost identical names.

 

[III-V]

Why would it be FUD.?.

Dare to give some proves first because if you have none its just plain deffamation, i posted a link that say so :

That only applies to the 8 series chipset, not the 9 series, which supports Broadwell.

Hfr article is dated 26/08/2013 , that s a little more than one month old, and the articles is not an interrogation, they says expressely in their title that :

We're in 2014, bud.

I'm very confused about people who talk about Intel having a current process lead. From what I can see, 22nm has far worse performance than 28nm in SOC form (bay trail fiasco) and only marginally better than 32nm sandy bridge in socket form.

22nm blows 28nm out of the water. The two processes aren't even remotely close. If what you were saying were true, Altera would not be doubling their performance by going to 14nm.

 

[Abwx]

That only applies to the 8 series chipset, not the 9 series, which supports Broadwell.

We re talking of Mboards, sure that the chipset is compatible, it s mentioned in the link i posted, but what about the current LGA 1150 MBs.?.

Can one buy a 1150 MB today, slap a i5 4670 in the waiting and then do a straight upgrade using a BDW LGA.?.

We're in 2014, bud.

Surely but that doesnt change what is said in said article.

 

[III-V]

Crashtech said 9 series.

 

[Abwx]

He said :

9-series LGA 1150 boards

Edit : That said you are free to think that he talked only of 9 series chipsets, everyone is entitled to have an opinion...

 

[III-V]

"9 series"

 

[TreVader]

That only applies to the 8 series chipset, not the 9 series, which supports Broadwell.

We're in 2014, bud.

22nm blows 28nm out of the water. The two processes aren't even remotely close. If what you were saying were true, Altera would not be doubling their performance by going to 14nm.

I hear this a lot, but I see just the opposite in benchmarks and in real world experience. I'm talking about 28nm vs 22nm SoC.


TSMC 20nm is a more advanced, more dense process and it's available now. I really hope Intel delivers with 14nm. I am planning on buying Broadwell-U immediately, just as I bought Sandy bridge and Haswell on release.

I only skipped ivy bridge, and I think we can all agree the 3770K was disappointing

 

[IntelUser2000]

How is Core M with the same amount of cache, much lower base and much lower turbo going to keep up with Haswell?

You can wish for Apple putting U chips and not making thinner and lighter.

The Core i7 5650U has 500MHz higher base frequency than the 4650U and while the top Turbo clock is 100MHz lower, the dual core max turbo is 3.1GHz for 5650U and 2.9GHz for 4650U which along with 5% perf/clock gains would result in 10-15% better performance.

I really hope Intel has kept a secret SKU for apple that will help core M compete with haswell (currently clocked 600MHz higher than 5Y70)

This isn't possible. Even set at 6W TDP the Core M won't be able to keep its top 2.6GHz clock sustained(maybe 2.3GHz), nevermind 3.2GHz. Even Intel's reference systems vary a fair bit in score. It goes as low as 2.48 to high as 2.8. I bet 2.4-2.5 is the actual steady-state performance.

22nm has far worse performance than 28nm in SOC form (bay trail fiasco)

Their mobile results have so far been disappointing but the CPU has been pretty competitive. GPU sucks because they've been underestimating competition for years, and they probably still do. Their process was always better for CPUs than others like GPUs anyway(high output with less density).

 

[TreVader]

You can wish for Apple putting U chips and not making thinner and lighter.

The Core i7 5650U has 500MHz higher base frequency than the 4650U and while the top Turbo clock is 100MHz lower, the dual core max turbo is 3.1GHz for 5650U and 2.9GHz for 4650U which along with 5% perf/clock gains would result in 10-15% better performance.



This isn't possible. Even set at 6W TDP the Core M won't be able to keep its top 2.6GHz clock sustained(maybe 2.3GHz), nevermind 3.2GHz.

Is intels ark updated with the 5xxx-U series (Broadwell-U)?


It's my personal opinion that apple will release an iPad-Pro or surface pro 3-like device with the 5Y70.


Are you saying the i7 - 5650U will be a 2.2/3.2 w 4MB cache? That would be a BIG bump in speed!

That's what I would love to see. Intel needs to give apple a reason why they shouldn't use A9/10/11 in anything bigger than an iPad.

 

[crashtech]

Series 9 Boards should support BDW-K.

http://www.bit-tech.net/hardware/motherboards/2014/05/11/the-intel-9-series-chipsets-examined/1

Confirmed from Intel.

Yeah, a closer reading of the available literature seems to confirm that it's the 8-series that has the problem, and that 9-series is the chosen platform for Broadwell-K. Still a little bit distrustful because the available info seems a bit scarce.

 

[Idontcare]

I only skipped ivy bridge, and I think we can all agree the 3770K was disappointing

May have disappointing but at least it was on schedule. For as delayed as the 3770K-equivalent Broadwell on desktop is going to be when it finally gets here, we ought to be expecting more of a tock-level performance bump than a mere tick-level performance bump that IB brought.