Skylake Coming H2 2015

[Pelea]

Intel confirmed Skylake will be released between July-December this year:

http://seekingalpha.com/article/282...-earnings-call-transcript?all=true&find=intel

since we're obviously gonna go straight from Haswell --> Skylake (Skipping a generation of improvements!), how do you guys think this will affect the following elements:

- Pricing of the models: How cheap will the 13 and 15 inch be? SSds have been getting cheaper, and if you compare 2015 to 2013, ssds will have almost dropped by roughly 50% in price by the end of the year.

- Graphics: Will the integrated graphics finally be strong enough for casual gaming?

- Minimum SSd storage: Will it be doubled to 256gb for the base model?

- new features: Wireless charging? etc..

-Battery life:?

-Thinness? just kidding, who the hell cares about how thin it is

Other:?

I want to buy a laptop that will last me 5 years or longer, and it looks like the skylake upgrade may be the last pc we'll need to buy for a long time..

 

[T'hain Esh Kelch]

Why do you think that a current generation model will not suit you for that many years?

 

[Cloudsurfer]

Still not convinced that Skylake will be in our MacBooks this year. Sure Intel promises it will not delay Skylake, but I find it hard to believe we will see both U and H models out before August in our computers. If Broadwell's launch is any indication we will see Skylake-Y released by August and then U and H maybe early next year. I think that is a much more likely scenario.

Doesn't mean that I wouldn't love a Skylake MBP before August though.

 

[Pelea]

Why do you think that a current generation model will not suit you for that many years?

The integrated graphics aren't good enough (base 13 or 15 inch model), and i'm looking for a 10h+ battery, as i know it'll only retain 60-80% of its original charge over time.

 

[mangotears]

I suspected as much, but I wouldn't draw too much from that.
Intel released their U suffixed processors just recently, which means those will most likely make it into the Air, and the 13" rMBP before back to school period.

I'm guessing that Skylake's release strategy will be reversed from the past couple years, with the higher TDP components coming out first. That'll allow them to keep pushing out their latest technology while still keeping a good cadence in targeting each segment of the market.

Best case scenario, we get 15" Skylake rMBP's before holiday season.

 

[Pelea]

I suspected as much, but I wouldn't draw too much from that.
Intel released their U suffixed processors just recently, which means those will most likely make it into the Air, and the 13" rMBP before back to school period.

I'm guessing that Skylake's release strategy will be reversed from the past couple years, with the higher TDP components coming out first. That'll allow them to keep pushing out their latest technology while still keeping a good cadence in targeting each segment of the market.

Best case scenario, we get 15" Skylake rMBP's before holiday season.

why would they split the release of the 13 and 15 inch models?

 

[mcarling]

Still not convinced that Skylake will be in our MacBooks this year. Sure Intel promises it will not delay Skylake, but I find it hard to believe we will see both U and H models out before August in our computers. If Broadwell's launch is any indication we will see Skylake-Y released by August and then U and H maybe early next year. I think that is a much more likely scenario.

Tock releases are not as spread out as Tick releases because Tick releases require replacing virtually all the equipment in the fabs (Intel of course do not shut down all the fabs at the same time) and Tock releases do not require shutting down the fabs. However, while it's clear that we'll see some Skylake CPUs shipping in H2 2015, it's not at all clear that we'll see Skylake-H CPUs (suitable for the MBPs) shipping in 2016.

Even if Skylake-H CPUs are due to ship in time for Apple to release Skylake MBPs in October 2015 (optimistic scenario), that does not exclude the possibility of Apple releasing Broadwell MBPs in, for example, June 2015.

 

[dusk007]

- Graphics: Will the integrated graphics finally be strong enough for casual gaming?

They are already if you consider dedicated like the 750M to be strong enough. Iris Pro is not much behind (plays at about similar settings fluidly) and Iris is still fast enough to play quite a range of games reasonably well.

Just because Intel says Skylake is coming H2/2015 doesn't mean anything. They always say H1 H2 and what they mean is like the last quarter they release a handful of chips and ship 90% of what you expect a half a year or more later.
I would not expect Skylake so soon. It will most likely be some time in 2016 for all the chips that actually matter. There will be a Broadwell MBP before that.

 

[Samuelsan2001]

Tock releases are not as spread out as Tick releases because Tick releases require replacing virtually all the equipment in the fabs (Intel of course do not shut down all the fabs at the same time) and Tock releases do not require shutting down the fabs. However, while it's clear that we'll see some Skylake CPUs shipping in H2 2015, it's not at all clear that we'll see Skylake-H CPUs (suitable for the MBPs) shipping in 2016.

Even if Skylake-H CPUs are due to ship in time for Apple to release Skylake MBPs in October 2015 (optimistic scenario), that does not exclude the possibility of Apple releasing Broadwell MBPs in, for example, June 2015.

This seems mush more likely apple upgrading all their rMBP's in june and then a complete update on all macbooks to skylake in march-april of 2016....

 

[mcarling]

Intel confirmed Skylake will be released between July-December this year

The linked conference call transcript does not say that at all. The Deutsche Bank analyst asked whether or not some Skylake parts will start shipping in H2 2015 and Intel's CEO and CFO both responded without giving a straight yes or no answer. They both said that Intel would not hold Skylake back just to sell more Broadwell parts. In other words, Skylake will ship when Skylake is ready. That's all they said. Absolutely no confirmation that any Skylake parts will ship in 2015.

 

[maflynn]

So why bother with Broadwell then, I've seen other news stories about Intel not delaying Skylake because Broadwell is delayed. Seems kind of silly, to rollout a chipset that will only be current for a few months.

 

[mcarling]

So why bother with Broadwell then, I've seen other news stories about Intel not delaying Skylake because Broadwell is delayed. Seems kind of silly, to rollout a chipset that will only be current for a few months.

The alternative would be shutting all the 14nm fabs down until Skylake is ready. They cannot produce Haswell parts.

 

[maflynn]

The alternative would be shutting all the 14nm fabs down until Skylake is ready. They cannot produce Haswell parts.

Broadwell is 14nm as well, so why not let broadwell live for 2015 and then produce Skylake instead of rolling out Broadwell in June, and then rolling out Skylake a month later (worst case scenario)

 

[mcarling]

Broadwell is 14nm as well, so why not let broadwell live for 2015 and then produce Skylake instead of rolling out Broadwell in June, and then rolling out Skylake a month later (worst case scenario)

It is more profitable for Intel to produce Skylake parts than Broadwell parts, for the reasons explained in the conference call (linked in the first post). Holding Skylake back after it's ready would not make any sense for Intel.

People seem to think there is an investment in Broadwell that needs to be amortized. There is not. There is an investment in the Haswell microarchitecture, which is amortized over Haswell and Broadwell. There is an investment in the 14nm process, which is amortized over Broadwell and Skylake. There is virtually zero independent investment in Broadwell. That's the beauty of Tick-Tock.

 

[leman]

It is more profitable for Intel to produce Skylake parts than Broadwell parts, for the reasons explained in the conference call (linked in the first post). Holding Skylake back after it's ready would not make any sense for Intel.

People seem to think there is an investment in Broadwell that needs to be amortized. There is not. There is an investment in the Haswell microarchitecture, which is amortized over Haswell and Broadwell. There is an investment in the 14nm process, which is amortized over Broadwell and Skylake. There is virtually zero independent investment in Broadwell. That's the beauty of Tick-Tock.

I was wondering about the same thing, thank you for your explanation, it makes sense. Then let's hope that the next rMBP will have a Skylake CPU + stacked DDR4 + V-NAND SSD + Maxwell gen3 GPU. Now that is something I would really consider upgrading to

 

[mcarling]

I was wondering about the same thing, thank you for your explanation, it makes sense. Then let's hope that the next rMBP will have a Skylake CPU + stacked DDR4 + V-NAND SSD + Maxwell gen3 GPU. Now that is something I would really consider upgrading to

Don't get your hopes up for Skylake plus a discrete GPU in any 13" or 15" MBP. I expect Apple will consider Iris Pro 7200 to be fast enough for the high-end 15" MBP.

 

[leman]

I still believe that they will keep a dGPU in the higher-end 15" (unless the 7200 is 3-4x performance increase over 5200). But nevertheless, the dGPU is not really important for me, its really just some additional gaming capability here and there. And if we should get some form of stacked RAM, the integrated graphics will also get a huge boost. I would certainly prioritise high-speed RAM over the dedicated graphics.

 

[maflynn]

There is virtually zero independent investment in Broadwell. That's the beauty of Tick-Tock.

What you call beauty, I call asinine (not you, intel). One thing is for sure, Apple and indirectly intel, is not going to get my money for any broadwell based computers. If they think that makes business sense to them, that's fine, but for my limited budget, I'd rather not buy a chipset that will be obsoleted in a matter of months.

On another thought, its possible that apple skips broadwell, or delays any sort of skylake rollout. Why update their MBP in July, and then intel releases Skylake shortly after that. They'll need to hold off until 2016 if they wish to stick with their annual roll out schedule. IF that's the case, they'll largely skip the skylake chipset.

OR they could just avoid Broadwell and go for Skylake with a fall update to their MBP line.

 

[mcarling]

What you call beauty, I call asinine (not you, intel). One thing is for sure, Apple and indirectly intel, is not going to get my money for any broadwell based computers. If they think that makes business sense to them, that's fine, but for my limited budget, I'd rather not buy a chipset that will be obsoleted in a matter of months.

If Skylake were to follow Broadwell by only one month (which didn't happen), I'm sure Intel's CFO would not have batted a single eyelash. Intel's customers (computer manufacturers) don't share your concern. They want the latest and greatest CPU of the week (at the desired price point). How long it stays the latest and greatest is not a significant issue for them.

On another thought, its possible that apple skips broadwell, or delays any sort of skylake rollout. Why update their MBP in July, and then intel releases Skylake shortly after that. They'll need to hold off until 2016 if they wish to stick with their annual roll out schedule. IF that's the case, they'll largely skip the skylake chipset.

OR they could just avoid Broadwell and go for Skylake with a fall update to their MBP line.

My guess is that Cannonlake will probably follow Skylake by 16 to 24 months, so I expect to see Broadwell MBPs in the June to October time range, followed by Skylake MBPs in early to mid 2016, followed by slightly updated Skylake MBPs (similar to the early 2013 and mid 2014 refreshes) probably in 2017, all before Cannonlake.

 

[plastictoy]

If Skylake were to follow Broadwell by only one month (which didn't happen), I'm sure Intel's CFO would not have batted a single eyelash. Intel's customers (computer manufacturers) don't share your concern. They want the latest and greatest CPU of the week (at the desired price point). How long it stays the latest and greatest is not a significant issue for them.

Intel's general endpoint customers who actually care about releases and advances and such are mainly enthusiasts on the desktop. Or they're enterprise (servers, Xeon) looking to reduce power yet not compromise in performance. So Skylake on desktop this year would satisfy them.

Mobile demand? Most people would be fine with the latest from last year. There's a reason they don't market laptops with Skylake and Broadwell attached to them and simply go by year of release: the codenames mean nothing to the average shopper. If half the people on here never read the forums or news, they'd be buying the current model of any Mac and not be disappointed with the performance.

By the way, manufacturers do care since Skylake and Broadwell require different platforms (1151 vs 1150). Dell was quoted in saying they wanted to overhaul their designs for Skylake instead of tweaking for a drop-in Broadwell. Makes sense and OEM brand would be pissed to invest in some Broadwell systems just to have a Z100 platform launch weeks later.

 

[maflynn]

Makes sense and OEM brand would be pissed to invest in some Broadwell systems just to have a Z100 platform launch weeks later.

Indeed, along with customers as well, which is my point. I suspect Dell, Apple and other major buyers have more information on Intel's roadmap then we do, still, the close proximity of the two releases is not good imo.

 

[plastictoy]

If anything, they'd probably just cancel mobile Broadwell-H and shift the anticipated high-end quads to Skylake. Release everything this fall or around that time, desktop to mobile to server on same platform. Less fuss, less consumer fuss even.

Broadwell U should hold just fine for the year and it gives Intel time to sell old stock while ramping up new stuff that possibly will be limited in capacity. Also gives manufacturers something to work with for now while the major refresh is simmering internally.

 

[Mr. Wonderful]

My guess is that Cannonlake will probably follow Skylake by 16 to 24 months, so I expect to see Broadwell MBPs in the June to October time range, followed by Skylake MBPs in early to mid 2016, followed by slightly updated Skylake MBPs (similar to the early 2013 and mid 2014 refreshes) probably in 2017, all before Cannonlake.

I agree with Cannonlake following Skylake by 16-24 months, but I have a feeling, even more so than Haswell, that Intel is aware that 10nm is going to take time, and that they'll release a "1.5"-type evolution of Skylake sometime in between, this time, all the way down to mobile.

 

[plastictoy]

I agree with Cannonlake following Skylake by 16-24 months, but I have a feeling, even more so than Haswell, that Intel is aware that 10nm is going to take time, and that they'll release a "1.5"-type evolution of Skylake sometime in between, this time, all the way down to mobile.

They can do incremental updates to Skylake without a problem. Bump the GPU, refine the architecture, bring some new features promised in the Skylake platform slides. Good for another year or two. Once they are on Skylake, this issue over Broadwell abandonment and two converging paths in the span of months will end.

Cannonlake is an issue of cost and yields, not technically getting there. I believe the biggest hurdle is after this with unknown 450mm fab ambitions and whatever material and technical advancement is needed to produce single-digit nanometer chips.

 

[Pelea]

it does seem weird to release the broadwell quad core processors at the same time as skylake.. they must skip either 1 half or the other..