Intel Developer Forum (IDF) 2013 on September 10-12, what can we expect?

[liahos1]

has anyone actually seen bay trail running and or gotten impressions?

 

[mikk]

when should we expect some real coverage of idf? there are a ton of of devices there and its where baytrail is making its splash and ive yet to see one impressions blog post in the blogosphere.

maybe tomorrow?

Wednesday is the day for Bay Trail.

As I recently pointed out in a different thread, fudzilla is usually wrong.
http://www.theinquirer.net/inquirer...ks-up-14nm-haswell-successor-broadwell-at-idf

Theinquirer is usually wrong and here once again. He said shipping by end of this year and we will see it in products next year.

http://www.youtube.com/watch?v=rcE06hzCzHI&feature=youtu.be&t=4m25s

 

[Khato]

Theinquirer is usually wrong and here once again. He said shipping by end of this year and we will see it in products next year.

http://www.youtube.com/watch?v=rcE06hzCzHI&feature=youtu.be&t=4m25s

Thanks for the video link. Guess I was incorrect in it being with respect to Baytrail... and correct in that Broadwell products are not going to be available by the end of this year I do wonder how a tech reporter translates shipping into 'first products available by the end of this year'... Considering that shipping doesn't even mean that they have final silicon that they can be manufacturing for profit, it just means that it's stable enough to ship to customers to begin their validation.

 

[Nothingness]

Thanks for the video link. Guess I was incorrect in it being with respect to Baytrail... and correct in that Broadwell products are not going to be available by the end of this year I do wonder how a tech reporter translates shipping into 'first products available by the end of this year'... Considering that shipping doesn't even mean that they have final silicon that they can be manufacturing for profit, it just means that it's stable enough to ship to customers to begin their validation.

That indeed would be crazy to expect Broadwell even early next year for end-users given that we only start seeing Haswell machines and that Intel stock is probably very large, especially given how sales are lower than expected. Add to that the competition in that segment is close to nil...

 

[tential]

So you're saying that in 6 months, my shiney new haswell machine will already have something newer and better than it?

Seems odd that Haswell ships in June/July (correct me on the months), but Broadwell will be out before the year ends.

30% battery increase too?

I wonder what the prototype laptops are looking like. How can they even start sellign broadwell based laptops now anyway? With all the stock people have of Haswell, it wouldn't even make sense to start selling it.

 

[erunion]

Theinquirer is usually wrong and here once again. He said shipping by end of this year and we will see it in products next year.

http://www.youtube.com/watch?v=rcE06hzCzHI&feature=youtu.be&t=4m25s

Thanks for finding that.

He says a few moments later that broadwell will be available next year. So nothing groundbreaking. Still its good that they are on track. There was some (baseless) speculation to the contrary. The sooner broadwell comes out the sooner we'll see Skylake....and hopefully more cores.

Random thought on IDF:
I'm excited for Bay Trail. Silvermont is Intel's biggest architecture redesign in a while. Would be nice to see another Conroe.

 

[Idontcare]

So you're saying that in 6 months, my shiney new haswell machine will already have something newer and better than it?

Seems odd that Haswell ships in June/July (correct me on the months), but Broadwell will be out before the year ends.

30% battery increase too?

I wonder what the prototype laptops are looking like. How can they even start sellign broadwell based laptops now anyway? With all the stock people have of Haswell, it wouldn't even make sense to start selling it.

The only way it makes sense is if it is a staggered launch with respect to tier.

Meaning broadwell comes out in 6 months but it is truly just the top 3 high-end mobile SKUs for the >$800 products, meanwhile the other 98% of the volume of products shipped will still be the mid and low-end haswell chips.

 

[Arachnotronic]

The only way it makes sense is if it is a staggered launch with respect to tier.

Meaning broadwell comes out in 6 months but it is truly just the top 3 high-end mobile SKUs for the >$800 products, meanwhile the other 98% of the volume of products shipped will still be the mid and low-end haswell chips.

Not quite. It is my expectation that an LGA based Broadwell labelled "Xeon E3" will be the first Broadwell out of the chute. Following that will likely be the Ultrabook oriented ones since this is where Broadwell will actually matter.

 

[NTMBK]

Clearly the CEO of intel has NO IDEA what's going on. He's just an idiot.

See, that's what I've been saying for years!

 

[mikk]

http://www.anandtech.com/show/7318/intel-demos-14nm-broadwell-up-to-30-lower-power-than-haswell
http://www.computerbase.de/news/2013-09/intels-broadwell-in-cinebench-bei-4.9-watt/

A bit more about Broadwell from IDF.

 

[Khato]

The Broadwell power improvement is indeed impressive. But there's the important question of whether that 30% improvement stays constant for the higher end SKUs or not. Not to mention that's only CPU - what kind of efficiency improvement are we going to see on graphics?

 

[mikk]

They just wanted to showcase what improvements from 14nm alone are expected. BDW has a new graphics architecture, any improvements from GPU aren't just coming from the 14nm shrink.

 

[bullzz]

$300 chrome laptops. this now makes $500 tablets expensive

http://www.theverge.com/2013/9/11/4720470/hps-colorful-chromebook-14-costs-299-this-holiday-season

 

[blackened23]

$300 chrome laptops. this is now makes $500 tablets expensive

http://www.theverge.com/2013/9/11/4720470/hps-colorful-chromebook-14-costs-299-this-holiday-season

I'm under the impression that Haswell is being used in the Chromebook Pixel which is certainly not 300$. Am I wrong in this? The original chromebook pixel was not a low end machine at all, although it was tied to the joke known as Chrome OS (worthless for anything except web use). It also had a high resolution retina type display and cost 1400$ IIRC.

edit: It appears Google themselves are making a chromebook pixel with a high end Haswell SKU, it will cost 1200$+. The regular chromebooks by acer and HP will use low end Haswell SKUs. I guess intel will take money where they can....

Chrome OS is still terrible, though, but who can complain for 300$.

 

[blackened23]

$300 chrome laptops. this is now makes $500 tablets expensive

http://www.theverge.com/2013/9/11/4720470/hps-colorful-chromebook-14-costs-299-this-holiday-season

Chrome OS, though. I would definitely take an ipad at 500$ or a Nexus 10 over a chromebook, as Chrome OS is extremely limited. So limited that I would put it in the same realm as Windows RT.

Essentially, Chrome OS is only good for rudimentary web functions; even Android and iOS have more functionality and applications than Chrome OS.

 

[Fjodor2001]

$300 chrome laptops. this now makes $500 tablets expensive

http://www.theverge.com/2013/9/11/4720470/hps-colorful-chromebook-14-costs-299-this-holiday-season

Some technical details on that HP Chromebook 14:

http://www.mobilegeeks.de/idf-2013-hands-on-mit-dem-neuen-hp-chromebook-14-mit-intel-haswell/

Acer also announced a Haswell based 11.6" Chromebook weighing under 1 kg:

http://www.mobilegeeks.de/acer-chromebook-hands-on-11-6-inch-intel-haswell-unter-1kg-video/

Note that both articles are in German, but you should be able to extract the tech specs anyway, otherwise Google translate it...

 

[Fjodor2001]

So IDF 2013 is over now. Were there no detailed presentations on the Broadwell CPU, including its architecture or the new generation iGPU it's supposed to have?

In that case isn't that strange? Last year there we're lots of detailed presentations on Haswell which e.g. AnandTech covered, like these ones:

http://www.anandtech.com/show/6264/intel-haswell-architecture-slides-idf-2012

http://www.anandtech.com/show/6263/intel-haswell-architecture-disclosure-live-blog

If there were no similar Broadwell presentations at IDF this year, what could be the explanation for that? Is Intel becoming more secretive and therefore does not reveal such info until closer to launch, or could there be some other reason?

 

[ShintaiDK]

So IDF 2013 is over now. Were there no detailed presentations on the Broadwell CPU, including its architecture or the new generation iGPU it's supposed to have?

In that case isn't that strange? Last year there we're lots of detailed presentations on Haswell which e.g. AnandTech covered, like these ones:

http://www.anandtech.com/show/6264/intel-haswell-architecture-slides-idf-2012

http://www.anandtech.com/show/6263/intel-haswell-architecture-disclosure-live-blog

If there were no similar Broadwell presentations at IDF this year, what could be the explanation for that? Is Intel becoming more secretive and therefore does not reveal such info until closer to launch, or could there be some other reason?

Plenty of technical documents to be found if you wish to find them. Anandtech seemed to focus more on Apples event.

You didnt see the chipset roadmaps on Anandtech for example, did you?

 

[ShintaiDK]

In terms of graphene, there was a Q/A session afterwards. And while there was great progress, it was still a few generations away. I assume that means 2019/2021.

It also means no flexible devices until that time.

This is what you will need for foldable computing and devices that you can bend and flex. Krzanich was very specific telling that Intel researchers are doing a lot of work on this field and that they saw some great progress. However it also said that product wise graphene is still a few generations away, that is what Krzanich said at the Q and A.
Graphene is still not cost effective solution, it’s good for prototype chips but not for full scale cost aware production. Until that time we have to stick to lithography and copper as the most price performance rated material. You won’t see any Atom or Core based flexible chips in this decade, at last this is how it looks from today’s point

 

[podspi]

In terms of graphene, there was a Q/A session afterwards. And while there was great progress, it was still a few generations away. I assume that means 2019/2021.

It also means no flexible devices until that time.

[/I]

When you consider tick-tock, a couple generations is only 4 years

Assuming that that is when they are ready to start implementing, yea '19/'21 sounds about right. If graphene is as good as man claim, we may look back and think 4 ghz is about as quaint as the single-digit mhz chips of yesteryear...

 

[Khato]

If there were no similar Broadwell presentations at IDF this year, what could be the explanation for that? Is Intel becoming more secretive and therefore does not reveal such info until closer to launch, or could there be some other reason?

Good question. I'd chalk it up to a combination of factors.

1. This IDF was primarily focused on Baytrail. Why spend too much time hyping up what's coming with specifics when they already have a leadership product in Haswell and now Baytrail as well?

2. An offshoot of the above, they revealed so much about Haswell early because they were behind in both graphics (to AMD) and power (to ARM) compared to the competition. It's a "don't buy from our competitors because what we're coming out with is going to beat them bloody" marketing ploy.

3. They most definitely are being more secretive! Especially with respect to their 14nm process. Far as I'm aware Intel has released no details with respect to their 14nm process, in comparison to 22nm where they published basic SRAM results back in 2009. It's natural for this to carry over to architectural details for the same reason - now that they have no reason to create hype for the next product refresh all that revealing information early does is tip their hand. And when the competition is capable of adjusting their product to some extent in 1-1.5 years that's a non-trivial issue.

 

[blackened23]

As there are multiple IDF dates throughout the year, i'm sure we'll hear more about Broadwell near the end of 2013. If it is indeed 30% more efficient than Haswell, that is definitely an impressive feat - and intel will obviously want to show it off.

 

[mikk]

So IDF 2013 is over now. Were there no detailed presentations on the Broadwell CPU, including its architecture or the new generation iGPU it's supposed to have?

In that case isn't that strange? Last year there we're lots of detailed presentations on Haswell which e.g. AnandTech covered, like these ones:

http://www.anandtech.com/show/6264/intel-haswell-architecture-slides-idf-2012

http://www.anandtech.com/show/6263/intel-haswell-architecture-disclosure-live-blog

If there were no similar Broadwell presentations at IDF this year, what could be the explanation for that? Is Intel becoming more secretive and therefore does not reveal such info until closer to launch, or could there be some other reason?

Because they had Bay Trail to showcase and furthermore Broadwell is coming H2 2014 and Haswell just launched 3 months ago. No surprise really.

 

[ShintaiDK]

As there are multiple IDF dates throughout the year, i'm sure we'll hear more about Broadwell near the end of 2013. If it is indeed 30% more efficient than Haswell, that is definitely an impressive feat - and intel will obviously want to show it off.

Correct. All the Haswell information came at IDF in april. Not september 2012. In september 2012 there was only this basicly:

More or less the same information we get now with Broadwell.

So in April 2014 there should be extensive information in terms of slides and pcitures.

 

[mikk]

That's not true. IDF fall 2012 had Haswell sessions.