Intel Broadwell Thread

[witeken]

"Intel’s 14nm Core M detailed: Broadwell delivers huge performance, dramatic power reduction"

http://www.extremetech.com/computin...creases-dramatically-better-power-consumption

Not sure about the "huge performance" though. They're talking about 2-3% CPU performance increase over Haswell, and 20-30% for the iGPU.

Just read the AnandTech coverage of Broadwell-Y...

BTW, IPC increase is "> 5%".

I think where Broadwell will shine instead is in low power (~4.5 W) chips. I.e. low power for the performance it provides in that TDP range. Focusing on that likely made Broadwell non-optimal for high performance desktop use.

Intel has 2 processes for 14nm. Broadwell-Y uses the process that is optimized for low power.

 

[witeken]

According to Chiploco, three Broadwell-Y CPUs have already been discontinued by Intel... prior to their actual release. This includes the 5Y10 in the Asus products announced recently. I can't help but wonder if something is going very wrong with Intel's new process... no one else can do 20nm and below with any kind of reliability, maybe Intel can't either?

Fake.

 

[JDG1980]

Fake.

The document in question is on Intel's own website. Should we believe that, or a random tweet from some guy?

 

[liahos1]

http://www.cpu-world.com/news_2014/2014070601_Release_schedule_of_Broadwell_processors.html

cpuworld mentioned this awhile back

"Later in the year Intel will start working on apparently newer Broadwell "Y" F-stepping, which will be in production in November. It is supposed to be ready to ship in the second half of December 2014 or in January 2015."

 

[Nothingness]

The document in question is on Intel's own website. Should we believe that, or a random tweet from some guy?

FPiednoel is working for Intel but has proved in the previous years that his tweets shouldn't be trusted.

 

[inf64]

FPiednoel is working for Intel but has proved in the previous years that his tweets shouldn't be trusted.

Yep agreed. He does work for intel but is disconnected from reality most of the time(last example is this very discontinuation notice we can see on intel's website).

 

[Homeles]

FPiednoel is working for Intel but has proved in the previous years that his tweets shouldn't be trusted.

He's kookoo for Cocoa Puffs; that's for sure.

 

[mikk]

The document in question is on Intel's own website. Should we believe that, or a random tweet from some guy?

Because a new stepping is coming and the current E0 stepping isn't needed anymore soon. Core-M isn't dead, it's just that some people are too stupid.

 

[liahos1]

http://techreport.com/news/27015/intel-reveals-core-m-specs-performance

article on core m, there is a live webcast at IFA with Kirk talking about Core M

 

[witeken]

FPiednoel is working for Intel but has proved in the previous years that his tweets shouldn't be trusted.

Yep agreed. He does work for intel but is disconnected from reality most of the time(last example is this very discontinuation notice we can see on intel's website).

The fake part obviously refers to the fact that Broadwell-Y won't go anywhere but in the stores by the end of the year, in contrary to what the sensational article suggests.

 

[witeken]

http://techreport.com/news/27015/intel-reveals-core-m-specs-performance

article on core m, there is a live webcast at IFA with Kirk talking about Core M

Awesome, Intel's forecast at their investor meeting was spot-on. If you do the math, you'll see that Broadwell-Y is 2.16X more dense.

Calculation: (1.3B / 0.96B) * (131mm² / 82mm²)

Also interesting to see that Intel claims their Gen8 is 47% faster GPU with 20% more EUs.

 

[Ajay]

I'm more interested in how the BW LGA Quads clock. Not real interested in these low power CPUs ATM. The density gain is impressive - just hope it doesn't kill clock rates on the high end.

 

[liahos1]

Awesome, Intel's forecast at their investor meeting was spot-on. If you do the math, you'll see that Broadwell-Y is 2.16X more dense.

Calculation: (1.3B / 0.96B) * (131mm² / 82mm²)

Also interesting to see that Intel claims their Gen8 is 47% faster GPU with 20% more EUs.

should be about 62% more dense than the a7 per mm^2 too.

1.3bln / 82 mm^ 2 vs 1bln / 102 mm^2

i just wonder how this thing is going to perform. they are comping it against a downclocked to 4.5w tdp i5-4302Y,

if this thing isnt faster than an a8 (20nm) or denver k1 (28nm) i am going to be depressed. they did say this was 3x faster cpu and 2x faster gpu than the "fastest tablets out there" or something along those lines. so vague..need benches

anybody know where we can get this presentation?

 

[witeken]

HT in Core M is also a good choice.

Also interesting to see that Core M never consumes more than 1W during following tasks:

You can clearly see how much power the screen consumes. I wonder when IGZO will become widespread.

Availability "late October".

 

[liahos1]

HT in Core M is also a good choice.

Also interesting to see that Core M never consumes more than 1W during following tasks:

You can clearly see how much power the screen consumes. I wonder when IGZO will become widespread.

Availability "late October".

how much more efficient is IGZO suppose to be?

 

[witeken]

IGZO = 2X power reduction.

IGZO display tech finally makes it to mass market: iPad Air now, high-res desktop display soon

 

[Piroko]

Anyone else noticed the absence of Apple?

 

[Arachnotronic]

Anyone else noticed the absence of Apple?

Intel would never pre-announce an Apple product.

Notice also no Samsung.

 

[kimmel]

Anyone else noticed the absence of Apple?

That's really nothing new. I believe Intel gets approval from each of the vendors to show that type of data on the slides. Apple likes to announce things themselves.

 

[Fjodor2001]

IGZO = 2X power reduction.

IGZO display tech finally makes it to mass market: iPad Air now, high-res desktop display soon

Yeah, at the CPU power consumption levels we're at (especially at idle or low workloads), the display accounts for the most of the total system power consumption. So if it can be reduced by 2x, that will make a huge impact.

 

[Fjodor2001]

I wonder how long the 4.5W TDP Intel Broadwell-M 5Y70 will be able to Turbo to 2.6 Ghz in a fanless Ultrabook? I'm a bit worried that it'll not be for s long as on e.g. 15W laptop CPUs or 77/84 W desktop CPUs. After all going from the base clock at 1.1 GHz to 2.6 GHz is a huge Turbo leap. Compare to e.g. typically 3.5->3.9 GHz on desktop CPUs, and then they even are having much higher TDP headroom. So are we talking about Turboing only for a few seconds perhaps on the 5Y70?

However if it will be able to do that for the complete boot up, during Windows updates, burst workloads on websites, etc, then I think it'll be a huge seller. Then it'll effectively be a 2.6 GHz CPU for the workloads that most people use the laptop for.

Sure, it won't be able to sustain 2.6 GHz when compiling large source code projects, encoding movies, or gaming. But are there any other typical major tasks it won't be able to handle at 2.6 GHz?

 

[Khato]

With this latest round of information regarding the Broadwell Core M being released ahead of IDF I begin to wonder what Intel has planned for IDF itself? Sure there will be a greater level of detail, but now two of the big items we were expecting to see at IDF - 14nm and Broadwell - have been revealed in advance. Do they have even bigger announcements to make during IDF itself? Or did they just want to get the big stuff out in advance so as to get better coverage given that Apple has taken to scheduling a conflict with IDF.

 

[witeken]

The IGP.

They're gradually giving more information. They showed 14nm's improvement on a test chip at IDF13. They gave an overview of Broadwell and 14nm a month ago. Now at IFA with multiple companies announcing designs with the chip, they're announcing the real-world benefits of Core M and the SKUs. At IDF we'll probably get a full architectural and 14nm disclosure. Maybe we'll get an official roadmap. We'll get Cherry Trail. We'll get a 10nm wafer.

 

[liahos1]

The IGP.

They're gradually giving more information. They showed 14nm's improvement on a test chip at IDF13. They gave an overview of Broadwell and 14nm a month ago. Now at IFA with multiple companies announcing designs with the chip, they're announcing the real-world benefits of Core M and the SKUs. At IDF we'll probably get a full architectural and 14nm disclosure. Maybe we'll get an official roadmap. We'll get Cherry Trail. We'll get a 10nm wafer.

hopefully we get a NDA release around the time of IDF where we get some real benches on a ffrd core m device. really would like to see some 3rd party ppl evaluate the sku.

its great they are getting this out ahead of sept 9th bc a lot of people are going to be focused on the apple event. sort of weird they do it on the same day.

 

[liahos1]

sort of off topic - apologies but i thought this was interesting

http://newsroom.intel.com/community...-to-develop-innovation-in-wearable-technology

i really like intel's go to market with wearables here. pitch technology to companies with established distribution/design/brands in the wearables category already vs smartphone oems building watches