Skylake/Broadwell Roadmap Update @Vr-zone

[Fjodor2001]

I'm keeping my expectations in check, but I'm really curious as to what Intel has done to Skylake.

Yes, me too. But we know it's the Haifa team that has designed Skylake. They've created some nice designs before, like Sandy Bridge.

So maybe it's just the Israeli based Intel team showing the american based Intel company how to design CPUs well...

 

[III-V]

You are correct. Interesting find

Graphics frequency is way high: 1.6 GHz.

 

[mikk]

Graphics frequency is way high: 1.6 GHz.

DDR3 1.6GHz and not GPU frequency.

 

[III-V]

DDR3 1.6GHz and not GPU frequency.

Yeah, just caught that.

 

[witeken]

Sounds eerily similar to the hype for broadwell, and we all know how that turned out.
I am at best cautiously optimistic for skylake.

Agreed. I was probably a bit too optimistic. If Intel would double the amount of FLOPS, it seems more likely they double the amount of slices.

Is this Desktop too, or just Mobile?

No FIVR saves power?

1) Desktop will be announced/launched at Computex (first Tock that won't get covered by Anand...).
2) FIVR is one of the things done to increase battery life, etc. http://www.ee.bgu.ac.il/~digismps/files/papers_2014/fivr_140M_intel.pdf (BTW, does any other company have FIVR?)

 

[frozentundra123456]

I was actually talking about the optimistic statements from intel, BTW, not anything you posted. They also pretty consistently "under-estimated" the extent of the delays.

 

[witeken]

I was actually talking about the optimistic statements from intel, BTW, not anything you posted. They also pretty consistently "under-estimated" the extent of the delays.

Sure, but I changed my opinion. Intel will have to prove the hard way that we should be "ecstatic" about Skylake. I guess they're a bit of a mix between optimism and pessimism, if you look at the spring roadmap for BDW-U (although I watched the IDF keynote and there he said early Q1), Skylake and 2016 release for Broxton (and SoFIA).

But yes they were indeed pretty optimistic about the yield learning, but then again, we're talking about Intel, who has the best yields in the industry ~ever with FinFETs that other people are heavily struggling with (Applied Materials said in its earnings that they expect the FF ramp to happen in H2), so they probably didn't think the yield learning would sort of asymptote at such a low yield, after which it would only increase very slowly.

 

[mavere]

So if I get a brand spanking new Skylake when its out, and load up a third party video player like MPC in Windows playing a video encoded with H.265 the Skylake CPU will use its new hardware capabilities without the need for any configuration?

Or are there any prerequisites? Like the player will need to be patched? Or that I need a certain windows version? Or a certain graphics driver?

For each OS, there's a system-wide abstraction layer for broadcasting hardware decoding support.

It's not guaranteed that software X will know to look for codec Y in the hardware decoder list before doing anything else, but if you have the standard "approved" playback setup on Windows (MPC-HC, LAVFilters, and friends), this is not a concern as long as you stay up-to-date.

(it's also not guaranteed that the OS layer will recognize any arbitrary codec. e.g. Windows DXVA2 has no concept of VP8/VP9)

 

[tential]

Broadwell being a "Dud" would only matter to me if it was going to have a full year on the market as well.
It's not though, Skylake is already on its way, even if Broadwell was a "Dud" it was still the best option on the market.
Broadwell would have had to have been a regression for me to actually be mad about it. Since it's not, doesn't bother me.

Skylake was always where I felt things would be the most promising anyway. I can't wait to see some concept laptop designs and to see the IGP performance.

I have no clue how fast IGPs are, but do you think Skylake's IGP would be faster than my old 9800M GTS in my laptop? I'd love to replace my laptop from 2008 (Penryn/9800MGTS), with Skylake.

 

[ame132]

I have no clue how fast IGPs are, but do you think Skylake's IGP would be faster than my old 9800M GTS in my laptop? I'd love to replace my laptop from 2008 (Penryn/9800MGTS), with Skylake.

I am 99% sure that haswell's iris pro 5200 is noticeably faster than 9800m gts

 

[mikk]

Another Update from Vr-zone with a more specific SKL-S date. Introduction August-October. Unlocked Skylake-S is coming indeed from the beginning.

http://chinese.vr-zone.com/143433/intel-skylake-would-announce-aug-15-02162015/

 

[witeken]

A bit later than I expected and some earlier roadmap showed. I wonder if it has to do with 14nm issues or 10nm issues. At this point, I also don't really understand BDW-K's place in all of this. Maybe BDW-K is for people who want Iris Pro?

 

[DrMrLordX]

That would seem to be the case. It could be an exciting product, or maybe something that shows us why Broadwell has almost entirely disappeared from Intel's desktop lineup. We should find out in . . . May? June?

 

[IntelUser2000]

Maybe BDW-K is for people who want Iris Pro?

It looks like it. I mean the K version on that slide says 4+2 which means GT2 graphics and will lose to the 4+3e Broadwell K in GPU but beat in CPU.

What probably happened is that they are seriously messing up with Broadwell, and K isn't turning out to be anything they expected. That means Skylake gets released as a K for non-iGPU gamers, and Broadwell K for those who want Intel's best iGPU for LGA builds.

Seriously though, unless Broadwell K has a big secret hiding and the iGPU is lot better than going from Haswell to Broadwell on the U parts, I doubt the adoption rates will be better than Haswell GT3e. 30% gain on current Iris Pro would end up as the last year's promised GT 650. Well, I guess they can always cut the price premium down to say, $20 more and people will go for it.

I'm not a bit impressed by Intel does nowadays at all. Remember that they said they would bring Atom on a new process every year? That never happened. And nowadays even Core is mediocre. I will tell you the ARM guys are doing better than them though. Galaxy S6 on a 14nm process 1H of this year? Tegra X1 with double GPU performance? That's amazing! Innovation on one side, stagnation in the other.

 

[mikk]

For me BDW-K with Gen8 is a no go when I can get Gen9 with HEVC Quicksync and all the new features from Gen9 for a cheaper price. And based on the Gen8 tests we might get a 20-30% speedbump over HSW 4+3e which isn't groundbreaking. Not to mention that SKL CPU cores are more interesting as well.

 

[Valantar]

I am 99% sure that haswell's iris pro 5200 is noticeably faster than 9800m gts

Yep. Iris 5100 should be too, although not by much. Videocardbenchmark.net.

 

[oldsome]

I am a bit confused. Will Skylake for desktop be released before Skylake for laptop?

 

[III-V]

I am a bit confused. Will Skylake for desktop be released before Skylake for laptop?

Sounds like it. Or at least, Skylake-S and Skylake-H (quad core notebook) will be out first.

For me BDW-K with Gen8 is a no go when I can get Gen9 with HEVC Quicksync and all the new features from Gen9 for a cheaper price. And based on the Gen8 tests we might get a 20-30% speedbump over HSW 4+3e which isn't groundbreaking. Not to mention that SKL CPU cores are more interesting as well.

BDW-K was supposedly cancelled, so yeah.

I'm not a bit impressed by Intel does nowadays at all. Remember that they said they would bring Atom on a new process every year? That never happened. And nowadays even Core is mediocre. I will tell you the ARM guys are doing better than them though. Galaxy S6 on a 14nm process 1H of this year? Tegra X1 with double GPU performance? That's amazing! Innovation on one side, stagnation in the other.

Tegra X1's GPU performance should not amaze you. Nvidia should have been serious about their mobile GPU uarch from the start.

 

[dahorns]

Sounds like it. Or at least, Skylake-S and Skylake-H (quad core notebook) will be out first.

Have I missed something? The roadmap in the OP clearly has Skylake-U out before Skylake for desktop. I don't recall seeing anything in this thread that would change that.

 

[mikk]

BDW-K was supposedly cancelled, so yeah.

BDW-K isn't cancelled.

 

[witeken]

Have I missed something? The roadmap in the OP clearly has Skylake-U out before Skylake for desktop. I don't recall seeing anything in this thread that would change that.

Skylake is now being plagued by the same delays as BDW and Atom.

Skylake-U/Y is delayed until (presumably early) Q4, and SKL-S isn't going to launch at Computex after all, but has been postponed by 2 months or so, which flies in the face of the Q4 earnings call.

Brian Krzanich - Chief Executive Officer
Well, this is Brian. Let me first kind of answer how we are looking at this and we are not going to slow Skylake down. We said it will be a second half of this year. I don’t want to slow it down because it brings a lot of innovation, a lot of new capability to this market.

But remember, they are on the same technology, the same piece of silicon, it’s the same factory. All we do is change the piece of glass in the scanner to get a different product. So there is not a change or revamp of our factories that needs to occur for this.

Stacy Smith - Chief Financial Officer
Yeah. That last point is important. They are both 14-nanometer products for us. So it doesn’t change our factory profile. And just generally, the faster we bring out new features and cool stuff to the market, the better off we are. So we are not planning to slow down Skylake, if that was at the heart of your question.

 

[mikk]

Skylake is now being plagued by the same delays as BDW and Atom.

BDW delay was much bigger, same for Cherry View. A 1-3 months delay is nothing special. Similar delays happened to HSW and IVB. 14nm delays were unusual big.

Skylake-U/Y is delayed until (presumably early) Q4

You don't have a source for that claim.

 

[witeken]

Of course I do..

http://fudzilla.com/news/notebooks/37015-notebook-skylake-coming-in-q4-2015
http://fudzilla.com/news/notebooks/37039-chromebooks-with-skylake-in-q4-15

I don't think the Skylake-S delay is 'nothing special'. Haswell will be in use for more than 2 years now on the desktop, so there's no reason to delay it. It was already in "ecstatic" health at IDF.

 

[krumme]

Is there reason to beliewe this is not just more intentional and a change of plans to reduce capex and variable cost than nessesary a delay because of process trouble?

 

[mikk]

Of course I do..

http://fudzilla.com/news/notebooks/37015-notebook-skylake-coming-in-q4-2015
http://fudzilla.com/news/notebooks/37039-chromebooks-with-skylake-in-q4-15

This is dubious. For example this part:

Intel plans to launch a few Skylake-U 15W TDP Core i7 and Core i5 with more to come in Q1 2016.

He forgot the time plan on those first ULT parts.

The company is commited to bring the Skylake in 2016

This guy can't even correctly differ between 2015 and 2016.

I don't think the Skylake-S delay is 'nothing special'.

It is. Look back to IVB and HSW. 1-3 months delay is business as usual. Comparing that to the huge BDW delay is just dumb. And even Q4 for SKL-ULT is earlier than expected. A month ago nobody here expected Skylake-ULT this year given that BDW-ULT just launched.