Intel Broadwell Thread

[witeken]

For the lack of a better thread:

http://semiengineering.com/manufacturing-bits-march-17/

In 2012, a startup called Zplasma came out of stealth mode and introduced its first technology—a next-generation power source for extreme ultraviolet (EUV) lithography.
(...)
"The status of the company now is exactly the status of the company three years ago: lab prototype proving (that) the breakthrough is real, exclusive license to the patents, and no capital with which anything can be built,” said Henry Berg, chief executive of Zplasma, in an e-mail.

“Having talked to everybody in the world with any interest in EUV, we have concluded that there is nobody to approach who will consider funding a new EUV source technology,” Berg said. “We would be happy to spring into action if this changed. The reaction we keep getting is: ‘This looks amazing. Wish we’d seen this earlier, but it is too late now.’”

Industry is in desperate need for EUV, but doesn't fund "amazing" technology. Would be interested in Idontcare's engineering perspective.

Edit: Another quote:

“The funny part is that when we started this dialogue in 2012, the most common reaction was: ‘You’re too late because we’ll be up and running later this year with LPP.’ Now, here we are three years later and EUV for HVM is as far away as it ever was. We would have shipped our source long ago, but we still haven’t been able to start the work,” he added.

 

[kimmel]

For the lack of a better thread:

http://semiengineering.com/manufacturing-bits-march-17/

Industry is in desperate need for EUV, but doesn't fund "amazing" technology. Would be interested in Idontcare's engineering perspective.

Edit: Another quote:

At first glance I would assume that if they really, truly, had as good of a technology as they claim in that article, then people would be lining up to fund them. That isn't happening so there probably is something more to the story.

 

[krumme]

I just did a bit a research, and some of the values I got contradict each other say from Broadwell to Haswell vs Ivy Bridge. It's quite strange. I didn't do a comprehensive analysis just because thats too much effort but, it's interesting to see the differences. Basically, from Sandy Bridge to Ivy Bridge there was a solid jump in performance, onward from there Haswell was barely any better, but Broadwell does look okay at least in comparison to Haswell vs. Ivy Bridge.

Keep in mind though that Haswell was 15W TDP with the PCH, whereas Ivy Bridge was 15W without the PCH (I think it was 18W or 20W total).

Hopefully Skylake is an Ivy Bridge like gain in performance, it's time for them to show up, because they have an ARMy coming after them at sub-10W, especially with A72/Custom and whatever Ares is.

I wouldn't hope for higher skylake performance any more. Performance is stagnating for good reasons but we got other things. Hw gave better battery life. And when samsung and qcom custom cores hit market giving competition (remember a72 is positioned as midrange), and with skylake reducing bom we will hopefully see lower prices. Ultrabook perf for all. That might be the big step and worth hoping for.

Ib ulv cpu perf is good as it is for single thread imo. I am sure Intel can easily add 2 cores if nessesary only adding little marginal cost. But hey if fanless is the most important then we are mostly there already.

 

[Vesku]

At first glance I would assume that if they really, truly, had as good of a technology as they claim in that article, then people would be lining up to fund them. That isn't happening so there probably is something more to the story.

There are only a few companies who could use that technology and if they have already invested 10s or 100s of millions of dollars into getting their current source to work it will be hard to drop it for this new offering even if it is promising. Which is exactly what they've said to this ZPlasma company: “We would be happy to spring into action if this changed. The reaction we keep getting is: ‘This looks amazing. Wish we’d seen this earlier, but it is too late now.’”

They might be interested in buying the patents but probably not at a price ZPlasma's investors would readily accept. This is actually true of many industries involving complex devices and technology, unless you show up with a 100% working solution there is quite a bit of reluctance. Even with a complete solution you may not be offered what it's worth since they know how difficult it would be to get the funding required to become a new competitor.

 

[danjw]

When are the quad-core mobile Broadwell processors due out? May?

Desktop and higher performance notebook processors have been pushed, yet again, from Q2 2015 to H2 2015. From what I have seen I wouldn't expect them before August, if then. In January they released two i7 mobile parts: Core i7-4720HQ and Core i7-4722HQ. These are 47 and 37 watt TDPs respectively. I doubt they would have refreshed these parts if they were intending to get Broadwell out real soon.

 

[alcoholbob]

Is it just me or are Apple MacBooks the only laptops out with iris pro 6100 IGPs? So if you want a windows laptop with a decent IGP right now you basically have to dual boot a Mac?

 

[coercitiv]

Is it just me or are Apple MacBooks the only laptops out with iris pro 6100 IGPs? So if you want a windows laptop with a decent IGP right now you basically have to dual boot a Mac?

How is that Iris Pro 6100 is the only decent IGP?

 

[Qwertilot]

Fastest IGP of the extant Broadwells I think? More to come of course and some decent Haswell IGP things.

I guess because Apple like very big IGP performance and have the ability to command a premium/enough volume and leverage over intel to actually get the things at affordable prices

 

[alcoholbob]

How is that Iris Pro 6100 is the only decent IGP?

Sorry - decent for me. I've been saying in this thread Id like a laptop that can match a HD 4600 on my 4770k before I upgrade since I play a lot of Diablo 3 and HD 4600 can average around 40fps in D3 at 1080p. I forgot in a long thread like this you have rewrite the same essay in every post since nobody reads further back than a page. Currently only MacBook Pros with HD 6100s qualify. I'd have to dual boot - what a pain.

 

[Dave2150]

Desktop and higher performance notebook processors have been pushed, yet again, from Q2 2015 to H2 2015. From what I have seen I wouldn't expect them before August, if then. In January they released two i7 mobile parts: Core i7-4720HQ and Core i7-4722HQ. These are 47 and 37 watt TDPs respectively. I doubt they would have refreshed these parts if they were intending to get Broadwell out real soon.

Could you link the official source from Intel where it says that the remaining Broadwell chips have been delayed from Q2 to 2H 2015 please?

 

[Fjodor2001]

Who cares about Broawell. It was an embarrassingly delayed set of SKUs that had to be released only to save the face of Intel. It'll have a ridiculously short life span. The attention is on Skylake now. Not sure how much better that'll be performance-wise though. Might add some nice wireless features though according to what Intel has communicated. Not sure who's interested in that however.

 

[III-V]

Might add some nice wireless features though according to what Intel has communicated. Not sure who's interested in that however.

Er, 90%+ of end users?

 

[crashtech]

Who cares about Broawell. It was an embarrassingly delayed set of SKUs that had to be released only to save the face of Intel. It'll have a ridiculously short life span. The attention is on Skylake now. Not sure how much better that'll be performance-wise though. Might add some nice wireless features though according to what Intel has communicated. Not sure who's interested in that however.

Why are you trolling? If you don't care about Broadwell, you can go away and leave this thread to the ones who DO care.

BUT, in the off chance your question is serious, I do have a mild curiosity about Broadwell, enough to participate and ask an honest question or two. So I guess that means I care. Add crashtech to your list.

 

[Fjodor2001]

Why are you trolling? If you don't care about Broadwell, you can go away and leave this thread to the ones who DO care.

BUT, in the off chance your question is serious, I do have a mild curiosity about Broadwell, enough to participate and ask an honest question or two. So I guess that means I care. Add crashtech to your list.

My point is that Broadwell will have an unusually short life-span, so it'll be a parenthesis in the Intel CPU history. I agree it might be interesting to discuss some technical aspects of it still though.

 

[Fjodor2001]

Er, 90%+ of end users?

What Skylake specific wireless features would that be?

 

[ninaholic37]

Why are you trolling? If you don't care about Broadwell, you can go away and leave this thread to the ones who DO care.

Considering that no one else replied to the post in 2 days, I just assumed everyone else agreed with Fjodor2001 and had nothing else to add to his assessment. Maybe you are the one who is in the minority? He is talking about Broadwell though so the post/opinion is on topic, either way.

 

[Phynaz]

Considering that no one else replied to the post in 2 days, I just assumed everyone else agreed with Fjodor2001 and had nothing else to add to his assessment. Maybe you are the one who is in the minority? He is talking about Broadwell though so the post/opinion is on topic, either way.

No, we didn't reply because we are refraining from feeding the trolls.

 

[Fjodor2001]

It's funny how the biggest trolls always accuse others of being trolls on this forum. I believe the psychological term for that is called Projection.

 

[bullzz]

agree with phynaz

broadwell 5770R (or whatever its called) is expected to have the same life span as 4770R. there is no info on skylake launching with edram sku

 

[ShintaiDK]

There is, 4+4e.

 

[Fjodor2001]

Yup, I believe 65W TDP Broadwell-K (5570R or whatever it'll be called) will be succeeded by Skylake S-3:

SKL-S-3: 4 cores, GT4e, DDR4 2133 MHz or DDR3L/DDR3L-RS 1600 MHz, 64 MB eDRAM, 35–65 W TDP.

 

[ShintaiDK]

It should be 128MB. Since Intels eDRAM is only made at 128MB.

With Broadwell it looks like this:

The wikipedia is an awful place to get information. The source of the 64MB is no less than the useless rubbish of gossip site called wccftech.

 

[Sweepr]

If there will a Skylake-K this year it will most likely be a 95W 4+2 SKU. I doubt we will see 4+4e SKUs till 2016 so Broadwell-K GT3e should have a fairly regular lifespan.

Quite a lot of people using Haswell Celeron/Pentium/Core i3's, massive amounts of DDR3 RAM or expensive LGA1150 motherboards have their reasons to be interested in Broadwell-K. Also some CPU applications should benefit from that huge eDRAM.

IMHO it's 1000x more exciting as a product than Kaveri Refresh and gives LGA1150 users flexibility to upgrade.

 

[Phynaz]

There is, 4+4e.

Codename overload in that slide.

 

[Fjodor2001]

If there will a Skylake-K this year it will most likely be a 95W 4+2 SKU. I doubt we will see 4+4e SKUs till 2016 so Broadwell-K GT3e should have a fairly regular lifespan.

Quite a lot of people using Haswell Celeron/Pentium/Core i3's, massive amounts of DDR3 RAM or expensive LGA1150 motherboards have their reasons to be interested in Broadwell-K. Also some CPU applications should benefit from that huge eDRAM.

IMHO it's 1000x more exciting as a product than Kaveri Refresh and gives LGA1150 users flexibility to upgrade.

Hmm... Broadwell-K is expected to be 65 W TDP, so how come Skylake-K would be 95 W TDP? Is there any source confirming that? :hmm:

From what I've seen the only 95 W TDP Skylake variant is SKL-S-2:
SKL-S-2: 4 cores, GT2, DDR4 2133 MHz or DDR3L/DDR3L-RS 1600 MHz, No eDRAM, 95 W TDP.

But it does not have any eDRAM... Is that the one you mean will be "Skylake-K"? Then what would they call SKL-S-3 which seem closer to Broadwell-K when looking at the tech specs?