Intel Broadwell Thread

[witeken]

Surely not. CHT after all delays is coming Q1-Q2 2015. You should better not expect Broxton for mid-2015. In the last Roadmap Intel already corrected it to end 2015, after the latest delays I don't think Broxton is coming 2015.

From the Q2 earnings call:

Then after that is, as Stacy said earlier, Broxton, which is our internal 14-nanometer product. That's targeted towards the mid to high level. And as we bring that into the second half of 2015 and into 2016, there will be various levels of integration on that.

It seems that Broxton might be somewhat affected by CT's delay, but since it is also produced on 14nm, it will not be delayed because of technical issues.

 

[Shyatic]

So all in all -- do you think this is a worthy upgrade to my i5 Sandy Bridge? I'm due for an upgrade next year... figure this might be a good way to go and donate my i5 to my wife as her desktop.

 

[Ajay]

wow that's really small. i wonder if intel should have kept die size ~100mm and used the extra real estate for more performance. piednoel cryptically laughed off denver k1's performance in a reply to a tweet yesterday (again). interested in seeing benchmarks for this.

Intel placed specific limitations on the TDP and and efficiency. The die size is the result of those (limits). So long as Intel pursues this strict efficiency benchmark, we will not be seeing very significant improvements in single threaded throughput. Fortunately, tocks have a bit more leeway to improve throughput a bit more at the expense of TDP.

I think we are all really interested in seeing the benchmark and power numbers for Broadwell; sadly, we'll have to wait a bit.

 

[Ajay]

So all in all -- do you think this is a worthy upgrade to my i5 Sandy Bridge? I'm due for an upgrade next year... figure this might be a good way to go and donate my i5 to my wife as her desktop.

IMHO, wait for Skylake.

 

[VengenceIsMine]

I wish the surface pro 3 I just bought had broadwell. More performance, less heat, less throttling.

Broadwell M is the first Intel CPU that will make the whole Surface concept/detachable tablet viable in terms of form factor, battery life and performance. It needs to be light enough to use as a tablet in one hand in an under 10" form factor in the consumer space or about 25% lighter with less thermal throttling in the current size for the heavier user. If you take something like the Surface 2 size and aspect ratio, run it on Broadwell M, get 8-9 hours battery and close to Surface Pro 3 perf then the concept might actually take off if Threshold/Universal apps finally build some OS momentum in the consumer space and if Intel isn't looking for massive $ for this part which I suspect that they will. My guess is that the new Microsoft CEO has had enough with $ losses in this space (look at how many hw projects have been canceled since he took over) and will cut down to 1 SKU which will then try to be too many things to too many people and be a muddle. Hopefully Lenovo or ASUS will make something interesting out of this.

 

[VengenceIsMine]

If rumours are true Core-M (Broadwell-Y) Turbo maxes out at 2.6GHz, which puts this 4.5W tablet / convertible chip firmly in Ultrabook territory when it comes to CPU performance (especially ST). Perfect chip for premium Windows / Android mobile devices later this year.

Hard to see how this can compete in the Android space pricewise. $200-300 CPU's are common enough in Wintel land but that's a MASSIVE price premium in Android land. Sure it'll smoke an Allwinner A80, Qualcomm Snapdragon 805 and probably whatever Apple's next gen is but you are talking 3x to 5x the price for the CPU/GPU alone and PCB costs will be much higher as well and that doesn't even begin to address the still massive PCB size & weight vs a top notch ARM based SOC tablet.

 

[witeken]

Hard to see how this can compete in the Android space pricewise. $200-300 CPU's are common enough in Wintel land but that's a MASSIVE price premium in Android land. Sure it'll smoke an Allwinner A80, Qualcomm Snapdragon 805 and probably whatever Apple's next gen is but you are talking 3x to 5x the price for the CPU/GPU alone

The answer is that it won't compete.

and PCB costs will be much higher as well and that doesn't even begin to address the still massive PCB size & weight vs a top notch ARM based SOC tablet.

Where do you base this statement on, source?

 

[TuxDave]

So all in all -- do you think this is a worthy upgrade to my i5 Sandy Bridge? I'm due for an upgrade next year... figure this might be a good way to go and donate my i5 to my wife as her desktop.

For desktops, I do the "upgrade when the things I like to do seem sluggish". I last did that when I bought a Core i5 750 for my gaming computer and it's still technically ok. As for laptops/notebooks/whatever, anything pre-Haswell is worth an upgrade in my opinion. 🙂

 

[Shyatic]

For desktops, I do the "upgrade when the things I like to do seem sluggish". I last did that when I bought a Core i5 750 for my gaming computer and it's still technically ok. As for laptops/notebooks/whatever, anything pre-Haswell is worth an upgrade in my opinion. 🙂

This would be for my desktop. I am trying to get a little into Twitch streaming so I'd like to either offload the processing to another PC (hence I have an upgrade so I use the old PC for that), or use one PC to cover both streaming and gaming and give the PC to my wife. Not sure she would want/need it though, but we'll see 🙂

 

[VengenceIsMine]

The answer is that it won't compete.


Where do you base this statement on, source?

If it doesn't compete at least somewhat, it simply won't sell in the Android market which like the PC market is viciously price sensitive. A $200 price premium and larger, heavier device is going to have a very limited market even with a big performance gain.

As for the PCB size/weight, are you serious? Take 1 look at a well designed PC PCB like the Surface Pro 3 PCB from a teardown like IFixit vs a current top of the line Android device like the Samsung Notes and you'll see a massively larger, more complex PCB with probably 3 to 4x the # of total chips and much larger size and weight and that's without even including the active cooling system. There is simply less integration in the PC side, has more RAM to support and it supports significantly more functionality & legacy standards. Even PC's based on current gen Bay Trail Atoms like the Dell Venue 8 are pretty massive compared to a Nexus 7 (2013) PCB much less an iPad Air, just browse a few teardowns and the difference is obvious without even going into BOM costs.

 

[witeken]

Guess what Intel did with Broadwell-Y?

XY reduced by 50%
Z reduced by 30%

Should've watched the webcast instead of comparing with Haswell-Y. You might then also hear that Broadwell-Y doesn't need active cooling. You say that the cost and weight will still be much higher, so I wonder if BDW-Y still won't be enough at all.

 

[dahorns]

If it doesn't compete at least somewhat, it simply won't sell in the Android market

I doubt Broadwell-M is intended for the Android market.

 

[kawi6rr]

you aren't the only one.

With no competition Intel has no reason to do more then the bare minimum.

 

[monstercameron]

With no competition Intel has no reason to do more then the bare minimum.

intel has tonnes of competition.

 

[VengenceIsMine]

I doubt Broadwell-M is intended for the Android market.

That was my point, it's not suited to it either technically due to form factor or by price.

 

[VengenceIsMine]

Guess what Intel did with Broadwell-Y?

XY reduced by 50%
Z reduced by 30%

Should've watched the webcast instead of comparing with Haswell-Y. You might then also hear that Broadwell-Y doesn't need active cooling. You say that the cost and weight will still be much higher, so I wonder if BDW-Y still won't be enough at all.

Those dimensions are still significantly larger than a higher end ARM chip and Intel still isn't doing enough to reduce the overall complexity and size of the typical Intel based PCB. Reducing the size of the packaging of the CPU/GPU & voltage regulation is only about 20-25% of the problem. Broadwell M based designs are still going to be larger & heavier than high end ARM based designs and when you throw in the price difference it's a steep hill to climb even with a significant performance advantage, though it's an open question how much of a performance advantage it will be over Apple's next SOC at TSMC 20nm which is likely to be it's nearest competitor on per core performance in mobile.

There is no doubt this is an improvement on the PC side over haswell but there will still be significant gap vs ARM based competitors on price and form factor, the real question is how much will the next generation A57 and custom core based ARM designs will close the performance gap since Intel is treading water on per core performance to pursue power gains again this product cycle.

 

[liahos1]

Those dimensions are still significantly larger than a higher end ARM chip and Intel still isn't doing enough to reduce the overall complexity and size of the typical Intel based PCB. Reducing the size of the packaging of the CPU/GPU & voltage regulation is only about 20-25% of the problem. Broadwell M based designs are still going to be larger & heavier than high end ARM based designs and when you throw in the price difference it's a steep hill to climb even with a significant performance advantage, though it's an open question how much of a performance advantage it will be over Apple's next SOC at TSMC 20nm which is likely to be it's nearest competitor on per core performance in mobile.

There is no doubt this is an improvement on the PC side over haswell but there will still be significant gap vs ARM based competitors on price and form factor, the real question is how much will the next generation A57 and custom core based ARM designs will close the performance gap since Intel is treading water on per core performance to pursue power gains again this product cycle.

the llama mountain reference design is 550 grams and 7.2mm thick.
I think u are splitting hairs here with regard to how much devices based on this chip are going to be different in size and weight vs arm tablets. i dont think you're going to get anything with equivalent performance, weight, thiness etc.

and then they have atom for smaller devices.

 

[dahorns]

Those dimensions are still significantly larger than a higher end ARM chip and Intel still isn't doing enough to reduce the overall complexity and size of the typical Intel based PCB. Reducing the size of the packaging of the CPU/GPU & voltage regulation is only about 20-25% of the problem. Broadwell M based designs are still going to be larger & heavier than high end ARM based designs and when you throw in the price difference it's a steep hill to climb even with a significant performance advantage, though it's an open question how much of a performance advantage it will be over Apple's next SOC at TSMC 20nm which is likely to be it's nearest competitor on per core performance in mobile.

There is no doubt this is an improvement on the PC side over haswell but there will still be significant gap vs ARM based competitors on price and form factor, the real question is how much will the next generation A57 and custom core based ARM designs will close the performance gap since Intel is treading water on per core performance to pursue power gains again this product cycle.

Intel would have severely misread the market and its competitors if anything from the ARM camp is even within shouting distance of Broadwell-M as to performance.

Broadwell-M is clearly intended for the high end convertible market, where no ARM processor currently lives. The Atom line is where Intel intends to compete with the ARMy. I'll be shocked if performance of 1) Cherry-Trail isn't fairly close to the ARM competition or 2) Broadwell-M isn't significantly ahead of both Cherry-Trail and the ARM group.

I just don't think Intel intends these to be overlapping market segments. Not yet at least. Nor do I think it likely that Intel has so severely underestimated the competition.

 

[liahos1]

Intel would have severely misread the market and its competitors if anything from the ARM camp is even within shouting distance of Broadwell-M as to performance.

Broadwell-M is clearly intended for the high end convertible market, where no ARM processor currently lives. The Atom line is where Intel intends to compete with the ARMy. I'll be shocked if performance of 1) Cherry-Trail isn't fairly close to the ARM competition or 2) Broadwell-M isn't significantly ahead of both Cherry-Trail and the ARM group.

I just don't think Intel intends these to be overlapping market segments. Not yet at least. Nor do I think it likely that Intel has so severely underestimated the competition.

agreed. these guys are basically blasting the low end of the china market. playing in the dirt with allwinner etc.

http://online.wsj.com/articles/for-...henzhen-holds-key-to-tablet-market-1405106599

 

[VengenceIsMine]

Intel would have severely misread the market and its competitors if anything from the ARM camp is even within shouting distance of Broadwell-M as to performance.

Broadwell-M is clearly intended for the high end convertible market, where no ARM processor currently lives. The Atom line is where Intel intends to compete with the ARMy. I'll be shocked if performance of 1) Cherry-Trail isn't fairly close to the ARM competition or 2) Broadwell-M isn't significantly ahead of both Cherry-Trail and the ARM group.

I just don't think Intel intends these to be overlapping market segments. Not yet at least. Nor do I think it likely that Intel has so severely underestimated the competition.

Which was my point in the first place, I was pointing out that Broadwell-M isn't going to be seen much in Android devices and it would face significant hurdles vs ARM based designs, especially in tablets 10 inches and smaller. Obviously Cherry Trail is intended for that segment. While Cherry Trail might keep pace on CPU perf (though it's mostly an improvement on the graphics side than CPU), given where Intel is starting in that market it needs to more than keep pace, it needs outright technical superiority to make market headway without just continuing to buy market share with contra revenue. Not to mention god knows when we will actually see Cherry Trail in volume due to the 14nm ramp up while there are massive amounts of ARM based SOC designs that will hit this fall that will be competing with Bay Trail on performance at some really low prices: A80, RK3288, MediaTek's big.LITTLE A7/A15s + all the A53/A57 stuff that will probably start to reach market before Cherry Trail in 2015.

Based on what we've seen so far, while Intel is obviously making serious headway on power efficiency with Broadwell, there are several other major factors impeding Intel in their various thin and light efforts in the high end segment where this chip is intended to compete (Windows tablets, convertibles etc): price, form factor and general weak demand for Microsoft's efforts in that area and Broadwell is only addressing power and partially addressing form factor. Based on that, I'd say it's hardly looks like it's enough to help energize the ultrabook/convertible segment which is what Intel would really like it to do.

 

[VengenceIsMine]

agreed. these guys are basically blasting the low end of the china market. playing in the dirt with allwinner etc.

http://online.wsj.com/articles/for-...henzhen-holds-key-to-tablet-market-1405106599

Of course but to get into the Shenzen whitebox tablet makers that's what it takes. Getting some market share and getting their supply chain straightened out so they can compete with the tablet chip makers is really all that Intel can do at this point since the smart phone market is out of reach until Intel can finally sort out a decent integrated LTE modem which is 2015 at the earliest and that will only be good enough for non western markets.

 

[VengenceIsMine]

the llama mountain reference design is 550 grams and 7.2mm thick.
I think u are splitting hairs here with regard to how much devices based on this chip are going to be different in size and weight vs arm tablets. i dont think you're going to get anything with equivalent performance, weight, thiness etc.

and then they have atom for smaller devices.

Prototypes have several luxuries unavailable to production units but it's a step in the right direction. The question is how much of that design is plastic and whether they used a single layer touch screen, unlikely you'd see that in a production 12" tablet until costs significantly improve.

It's really, really not splitting hairs. Look at the lengths that Microsoft is going to with Surface, using magnesium etc. Apple can use much cheaper aluminum and still come out way lighter and in large tablets, that weight is really really key. Look at any review of Surface 3 or almost any 10' PC tablet and they all complain about the weight making it undesirable for use as a tablet. Considering that those devices are being positioned as a PC + a tablet, the fail as a one handed tablet is not trivial, those 100-200 grams in extra weight can be huge in terms of usability. It's not the only factor but it's anything but trivial.

Surface Pro 3:

IPad Air

 

[liahos1]

but that is comparing a 15W tdp part and all the associated parts needed for a real laptop versus a fanless tablet.

https://lazure2.files.wordpress.com/2014/04/image59.png

that is a baytrail entry pcb

http://cdn-static.zdnet.com/i/r/sto...erboard-620x438.jpg?hash=AGWzZGt5BJ&upscale=1

thats a core m pcb? clearly its still more complicated but but much less complicated than the right comp for it which i guess would be the mac air - pcb also in shot.

if they can build something with between i3/i5 levels of performance that is fanless and weights less than 600-700 grams. and is thinner than a surface pro 3 i think we start to get to diminishing returns

 

[Sweepr]

Broadwell article at ArsTechnica and some new slides:

http://arstechnica.com/gadgets/2014...-intels-low-power-core-m-and-its-14nm-process

 

[Homeles]

For desktops, I do the "upgrade when the things I like to do seem sluggish". I last did that when I bought a Core i5 750 for my gaming computer and it's still technically ok. As for laptops/notebooks/whatever, anything pre-Haswell is worth an upgrade in my opinion. 🙂

Unfortunately, it appears that being on SATA2 is holding me back quite a bit, so I'm itching for an upgrade, even though I am more than fine with my i5 750 as well.

I totally agree with you on laptops though. Haswell's big enough of an upgrade, even for Ivy Bridge users. Obviously, most people wouldn't be down to buy a new laptop only one year later, but there would actually be tangible gains for them to do so, which generally isn't the case.

At this point though, I think it'd be better waiting for Broadwell. Waiting 6(?) months for the kind of battery life and graphics performance it'll be bringing should be fairly reasonable. I'm also a bit pessimistic when it comes to Skylake for mobile, as the loss of the IVR is a big bummer.