Forget Intel Broadwell, Skylake On the Way

Probably not in your case... The Broadwell H has been delayed and delayed and go knows when it will come out... As for Skylake H, I wouldn't expect anything anytime soon.. So your best sticking with your P651SE...

 

Usually when people refer to battery life in a laptop, they refer to usable battery life. A cell phone sure you might be more interested in sleep time but laptop it's all about getting through a day off light use without having to plug in.

 

Actually it'll be different with Us and Ys which will go into Tablets and 2-in-1s. The new devices with Us and Ys wake up in <1s. That's fast enough so you can press the power button or close the lid between classes or meetings. Even with a much longer wake up mode on my Ivy Bridge Ultrabook and 6 hours of battery life I've gone for 2-3 days between charges. Now take the Haswell+ generation which depending on the device brings it to 9-11 hours.

Think about a scenario where you are only using 20% of the battery on active usage per day. Theoretically, you'd get it to charge it on the end of the fourth day. Now let's compare a device with 11 day standby vs one with 55 day standby.

11 day standby day 1: 20% active, (1/11)9.1% standby power = 70.9%
day 2: 41.8%%
day 3: 12.7%

55 day standby day 1: 20% active, (1/55)1.8% standby power = 78.2%
day 2: 56.4%
day 3: 34.5%
day 4: 12.7%

If I am going to buy a near $1000 device that I will keep for next few years $20-30 difference for that benefit is worth it IMO(of course in reality manufacturers might charge additional $50-100). Also, having an efficient component is always there, but batteries will degrade as time goes on, making the difference more important.

 

For what its worth, I've noticed the two Intel CPU design teams have fundamentally different design goals that favor certain parts of the market:

(I am skipping the ticks as they are not "architectural changes")
Haifa: Responsible for Core uArch(Conroe/Merom), Sandy Bridge, Skylake
Design goals-
*Microarchitectural performance improvements
*Core-level power management improvements
*Favors more for "PC" ie., Laptop, Desktop

Hillsboro Oregon: Responsible for Nehalem, Haswell
Design goals-
*System level performance improvements(QPI, IMC, Hyperthreading)
*System level power management improvements(PM chip on Nehalem, new improved power state on Haswell)
*Scalability features(easy cut and paste dies for making numerous die variations, and multiprocessor support)
*Favors Server and Workstation

You'll see some oustanding examples of "favoring". Core uArch chips were fantastic for PCs, but still lacked somewhat on servers/workstations. Nehalems were "ehh" because they brought multi-threading improvements which back then were not too relevant, but servers/workstations got HUGE gains. As a company as a whole they need both teams since their strengths are different but not all customers will be happy at the same time.

I am not sure if we'll see Sandy Bridge level perf/clock gains but Skylake should end up better than what Haswell brought.

 

If standby time is what they are focusing on then whoever did their market research should be fired immediately. If you surveyed users I can guarantee well under 1% would state longer standby time as a concern.

 

Agreed.. Who the hell cares about standby time? I use my notebooks every day and in an extreme case, my ultrabook get's used once a week..

 

Right. I usually turn my ultrabook OFF anyhow when not in use. It boots in like 8 seconds so who cares...

 

Celerons coming in Q4 15

 

Almost April in a day or so, computex just 2 months away, first week of June. Windows 10 a matter of months away. No doubt leaks and real info about to emerge. It often said there is something better around the corner, but at this stage big leap from mostly 4th gen quads to 6th gen quads is something to hard to ignore. There are those that say, Skylake will not a big increase in performance, although will be a big increase in efficiency for mobile platform there is no doubt about that. Many vendors will be looking at releasing new tech with new OS at the same or similar time.

 

Intel’s Broadwell to Skylake uArch Transition Will Be As Big As Prescott to Conroe

 

Haha...yeah I'll believe it when I see it...

Oh wait, WCCFTech? Nah they lit.

 

Friend, friend, friend...are you reading the links I post? I don't think so.

By the time Skylake releases (Q3 '15 at the earliest), LPDDR3 vs DDR3L price surcharge will be $2 for 4GB-equipped systems and $4 for 8GB-equipped systems.

I'm amazed you think your argument holds any strength...$4. The standby battery life gains are not useful every day, but see the post by "IntelUser": $4 is paltry.

Regarding shutting your laptop down: of course, you can just shut it down. But, then you have to manually save all open work and then re-open it all back. Why not just close the lid and resume where you where?

Laptops are becoming more like tablets. Who wouldn't want the battery life of a tablet in their laptop?

Not a "big deal", but it's one of the nice things about Skylake that I'm excited about.

LOL, Intel forcing them helps nobody...then why did they do it? Why did Microsoft include it in the Surface Pro 3? Why did Apple include in the MacBook Air and the newest 2015 Macbooks? Why?

Intel certainly didn't force Apple and we all know Apple chases those crazy high profit margins. Why would they pay a premium for LPDDR3?

The makers of LPDDR3? You're kidding...you must be kidding. The price of LPDDR3 has been dropping for months: not sure they're the big winners here, lol.

I never said standby battery life was as important as active battery life, lol. Who said that?

1) I wrote "battery savings" in just my first post (where LPDDR3 was just a minor point). Your first reply to me actually quotes my 2nd post where I go into detail about standby battery life, haha.

2) I mentioned the 19% increase in efficiency in active use just once (in my 2nd post)....this whole debate has been about standby life. Not sure why you think anyone cares about that 19%?

3) Again, nobody is claiming that 19% is worthwhile...maybe it's just the strawman you found in the field, heh.

4) Old enough to read all the posts, I think. And that remark applied to the posters above you, not you, mate. No worries, you're safe,

5) Your excitement is your own business (in my opinion).

6) Well, you'll likely be enjoying that LPDDR3, too, if you get a Skylake laptop. We can be LPDDR3 buddies!

That's some pretty interesting information; I didn't know the Conroe-team was designing Skylake.

But, when Conroe released, AMD had lit a fire under Intel. Back then, if you wanted performance, you did not buy Intel. Intel is in a far more comfortable market position now. Why would they want to increase IPC if they don't have to? I mean, all the funding for R&D in a market where they clearly dominate and have been dominating for the better part of a decade?

I feel, if the performance was clearly better, Intel would've decreased clocks and lowered power consumption and at least would have bragged about that at IDF '14. They were bragging about much less with Haswell and Broadwell.

Who said standby battery life is what Intel is focusing on? Who is this crazy person? What kind of ignorant user would post such a baseless fact? HTWingNut, please, quote this person for us!

It's a bit ironic how you mention that your Ultrabook, in an extreme case, gets used once in a week. That's....well.....kind of.....well, really, exactly....uh, never mind. I'm sure you're following our discussion.

 

I really think Intel and the rest of the PC market in general will have a hard time selling PCs to the general market (excluding enthusiasts that need the latest and greatest) due to the immense success that was Sandy/Ivy bridge. Processing up to this point has been more than enough for the things most people would do, with fairly decent usable battery life (~5 hours). People have no reason to upgrade, which is why I feel Intel has been artificially delaying the release of their new processors. The only time I see people upgrading, is when UHD becomes mainstream (meaning hardware AND content), where the intel HD3000/4000 would struggle to run I would imagine.

 

LOL... I always shut down my computers after use.. The standby arguement doesn't apply to me hence..

 

A one time cost of $4 or $400 over the lifetime of a system has always proven useful and cost effective if the benefits are realized every day of the lifecycle of the system in question.

Even if the big 'selling point' is not used in a specific workflow (I too don't use stand by or sleep...), it still overflows into other areas that can benefit almost as much (primarily less heat - killer of anything electronic) and therefore still makes it a worthwhile checkpoint when buying a new system. Especially when both options are in front of you and all you have to do is make the right choice.

The clients that make their choice based on pure $$ savings have invariably regretted it over the past 4 decades. Even if they don't admit it publicly (the proof? they buy the recommended system setup the next time...).

The clients that take my recommendations to heart seem to better understand the fact that a system is based on the harmonious interactions of all the discrete components. And appreciate the fact that a single one time surcharge for a 'balanced' setup is more than made up over the lifecycle of the system with reliable, dependable and consistent performance and availability over that time period.

When budget/$$$$ savings really count; I always suggest to forego warranties and instead buy better hardware instead. Same or lower cost in the end (decade average); much higher compute performance per year.

 

After discovering Skylake is BGA only... I won't shed a tear if it brings no substantial performance improvements over Haswell.

 

Skylake, Broadwell, everything from Intel in the future is unfortunately BGA.
We have lost the war. Intel and OEMs won

 

Progress requires some sacrifice. BGA only? So what? I'm sure there will be a Skylake cpu offering that matches and exceeds the performance and features of all previous platform for almost everyone.

The way the cpu is mechanically coupled to the M/B is inconsequential to the performance the platform can deliver.

Yeah; it will make upgrading or replacing a CPU problematic. But there really is no reason to do that anymore at all (and for me, not for more than a decade or more).

This is my view on notebooks. On desktops; I'm sure the moaning about BGA is more than premature.

 

1-2 years down the line. The CPU dies. Or overclocking kills the CPU earlier which warranty doesnt cover.
No problem, I just use the warranty. Wait, I forgot, it have expired.
Fine, I just buy a new CPU. Crap, I forgot, its soldered.
Just have to dish out a new complete motherboard and a new CPU soldered with it then. Oooops, Nvidia also solder the GPUs now.
Motherboard plus CPU plus GPU ends up costing more than a new notebook does.
Money in the bank for OEMs and everyone involved in making the notebook.
Customer lose, overclocking enthusiasts lose interest and move on to desktops.
All is left is hipsters that doesnt use the hardware much and only mediocre hardware is made for notebooks because enthusiasts and power users have moved on and everyone else is happy with any dirt Intel come up with. This have already begun with 4980HQ. A let down compared to 4940MX

 

I can't remember the last time that mine or my client's cpu's dying. This must be over 20 years or more now.

Overclock? Why? More power, more heat, more noise, less reliability and bigger issues (blue screens, etc.) over and above the normal O/S glitches. Not to mention that the first time a component dies, it has already negated any speed increase it offered by the downtime... And all this for ~20% or less performance increase. I don't get it.

Use components to spec's. Push them even to their max limits. But overclocking and wanting to overclock is like wanting the muscle cars of the 1960's today... sure, with a hammer, a wire and a screwdriver you could bring it back to life and depending on the parts installed, it could still teach a thing or two (but not all things) to modern equivalents. But the point is; you still have a 1960's car. Even if you think you have a 2015 'anything' equivalent.

Computers are cheap enough and powerful enough that worrying about overclocking is going the way of the dodo.

To get 'X' amount of performance you need 'y' and 'z' amount of platform and cpu. The total cost is the same (when you include your time). If you need more performance today, sell and buy what you need.

If you think you can get more performance by buying the lowest end platform and upgrading it piece by piece and still think you're saving money. Uh, okay... You're not. But, if it helps, keep thinking that anyway.

As a hobby, I can see the frustration. As a tool (which is all computers are) it is a simple decision to buy more/bigger/better as your compute needs change. And, if you're buying so often that you think you're being ripped off by manufacturers; the fault is yours not theirs. Define your needs and proceed accordingly.

But claiming that we need socketed cpu's because they die on us... well, that logic will get you nowhere.


(If you don't know what you're doing; learn. If you kill a few cpu's learning; that is the cost of educating yourself).

(And once you learn this 'skill'? You'll at least have the memories to share in the future, right)?

 

It's OK if you like underpowered low TDP spec'ed parts that throttle like crazy. Even if you CPU doesn't die you never have the option to upgrade, or for enthusiasts, push the limits. GPU's are a lot more subject to failure, and having those coupled with the CPU embedded means you have to replace CPU or GPU. So your failure mode includes typical motherboard, CPU, or GPU. Three times typically what you'd have at likely five to six times the cost.
I've had CPU's die on me. I've had GPU's die on me. I've had motherboard components die on me. For a $300 laptop maybe it doesn't matter, but for a $2500-$3000 high performance one, it matters.

 

And what about the small custom houses that offer multiple CPU options, my guess is now they are going to be limited as well. IMHO supporting BGA only is like saying here is the OS and the system is now bound to it, you can run nothing else. As we bind the systems we at the same time take away options. This is a bad thing for consumers and a great thing for manufacturers.

 

BGA is an unfortunate shift due to people buying hardware that seem cheaper in the short run(oooh, $50 cheaper) but more expensive if it breaks down. Also it has an advantage in more portable systems as well.

I wish people make upgradeable laptops. Unfortunately business trumps everything else.

Not necessarily. Lot of the new upgrades are due to replacement cycles - of their previous CPUs. Intel got big sales boost in one of the more desirable CPUs in their history - Sandy Bridge. Everyone loved it from Thin & Lights to Servers.

AMD is pretty irrelevant to them but now they got a much more formidable competitor, and that's based on ARM. They can't make the Atom performance too high if Core chip stagnates. They can't make Atom performance bad if they want mobile shares. Increasing IPC improves perf/watt which improves their entire lineup.

They don't do it because they can't do it.

 

Not going to answer these individually (the rhetorical questions will remain rhetorical).

The funny thing to me is that people want progress but they want legacy 'stuff' too.

Nobody is going to open the doors to their business and invite you on an all you can grab in 60 seconds mad dash through inventory.

Nobody will be motivated to push for that progress if it means it will put them out of business.

Nothing will progress if it means choosing to remain stuck at the place we happen to be at now.

BGA (now) and whatever else is coming down the pike is needed to take us to the next step, ad infinitum.

The next step of performance. The next step of low power usage. The next step of connectivity.

Will consumers need to adapt? Will third party providers need to adapt? Will the company that makes these changes need to adapt?

In all cases; yes.

Those that don't will be using circa 'xxxx's' platforms as long as possible and cursing the much better alternatives already available (at whatever time line you want to discuss). Or, they'll be out of business. Or, they'll take the new options and really show us what it means to innovate.

Whatever happens will be tempered by how precisely that 'next step' matches the needs of actual users. We are not blindly led down a dead end alley. We have our senses and our brains to see what is best for us. As I mentioned before; computers are simply a tool. Unless someone's hobby is OC'ing, that is a dead end pursuit for 99.9999999999% of users.

Getting the most out of your $$$$, your hardware and your time does not have to involve pushing your current system past it's spec's. One can simply choose to do so because things in the universe at this time have lined up so.

We can reminisce about great moments in the past. But that doesn't negate the great tech we all enjoy today and each future day to come.

Replaceable core components are great in theory. They have simply outgrown their usefulness and cost effectiveness when final performance/productivity per dollar is the goal.

I don't buy systems that overheat and/or throttle (get lost fruity company products...). I don't buy $5K+ systems that are undependable (thank you ThinkPAD's). And I don't buy into the notion that user replaceable is the better product in the long term (not when I'm chasing maximum productivity over time).

And as much as people complain about Intel's compute stagnation (lol...) over the last few years; progress has been made.

A client bought a Broadwell platform with an M.2 (SATA) SSD. Blows away almost any i7 IB based platform I have ever used (w/regards to 'snappiness') and yes, those too have SSD's installed. The 'unbelievable part'? The Broadwell platform is an i3 (shock!). Can't wait to play with an i7 QC based Broadwell (or better yet, Skylake) platform.

What! No sockets, just BGA? No problem. When/if I find the platform it is based on lacking, a cheap alternative will be waiting at the corner Staples store. But needing to fix it because of the cpu/gpu dying? Not in the last few decades...

This goes back to my recommendations to always buy the most hardware (and most current) you can afford. Given; O/S's don't require less compute power over time. Software the same. Users, even more so. Even the lowly browser can bring some systems to their knees. Buy the most balanced system you can afford; it is never wasted.

If a new system doesn't wow me (over and above what I previously had/have) it is returned, promptly. I don't have to consider potential/future upgrades to decide. The decision is made right there. Sideways upgrades are never worth it. Ever. And OC'ing and/or upgrading an aging cpu is almost always a sideways 'upgrade'.

What wow's me on some of the newer platforms? As already mentioned; the snappiness. The low power usage. The light(er) builds. The long(er) battery run times. The quietness of running the system flat out. The dependability of the newer systems because of all the things listed here.

What don't I care about? Whether or not the core components are replaceable (cpu and/or gpu). I ensure that these are at the levels or above that I need at time of purchase for the lifecycle of the system.

As long as I can upgrade the O/S at will, have easy access to the storage subsystem components and can max out a platform's RAM , everything else is just superfluous to the question I have of every new possible system: will this tool give me better overall performance (compute power and/or battery life and/or weight, etc.) than anything else I have/own today.

 

BGA has been around like forever, so who is talking legacy here with PGA vs. BGA? I have no issue with M.2, again an interface board. And it is fine you do not care about replacements but a bunch of us do and in the market majority rules apparently except in the case of Intel.

 

The issue is not PGA or BGA here (specifically; re-read my post).

It is not a matter of caring about replacements. These can always be had (one way or another). It is a matter of choosing the most logical course forward. Right now, BGA is that way and it doesn't matter how long it has been around before.

Majority rule is never a reason or cause for excellence. Neither do the majority want progress at any given time, if it means change.

 

Majority demand usually rules the marketplace, I never said it was good just the way it is. I mention BGA's time frame because you brought up the point of PGA being legacy, that is not the case. I am the first all out for change, where an actual improvement (on the consumers end) is involved.

 

Intel at IDF today:


Guess we should have it soon. Yay

http://www.eteknix.com/intel-skylak...face-recognition-with-realsense-3d-technology


Update #2:

http://www.fudzilla.com/news/notebooks/37472-skylake-15w-mobile-skus-getting-iris-graphics

 

Still nothing about the desktop processors... I honestly can't give a damm about the 15W parts...

 

Patience grasshopper.

 

Skylake launch at IDF in San Francisco.

 

Question is when is IDF in San Francisco this year? is it going to be June like last year or September?

 

http://hexus.net/tech/news/cpu/80786-intel-skylake-s-desktop-cpus-expected-idf-2015-august/

August 15.

 

OK, Just to clear up some confusion on my part, can we expect to see skylake notebooks in the 3rd quarter or 4th quarter? or should we be expecting any skylake notebooks at all this year?

Thanks

 

Your average consumer isn't likely to open up their computer. Nevertheless replace any components. And manufacturers don't trust people with modifying their hardware, because they might get blamed for it and thus potentially have their brand tarnished.

A friend bought a "gaming" desktop with a GT 610 a while ago, and hasn't opened the case ever since. Mind you that the GT 610 graphics card is equivalent to a GT 610M.

If there's no incentive from either consumers or manufacturers to modify laptop hardware, then there's no point for manufacturers to maintain upgrade-ability other than customizable laptop models (before they're shipped out).

 

Your guess is as good as anyone else's, but with all indications, that is the hope...

 

Yes, it's all just a guess at this point as to when Skylake will actually be available.

So - who wants to start a pool as to whether or not we'll even see mobile Broadwell quads?

 

That is a bet that is a sure 'lose'.

 

tilleroftheearth, progress is progress, and BGA is BGA. There is nothing progressive about BGA, it is just a way to maximize profits for manufacturers (and repairmen, consequently) by ripping off customers, compared to PGA that is. You are not praising progress here. You are praising the fact you (and everybody else) were ripped off. I fail to understand this overoptimistic enthusiasm.

 

Personally, I wouldn't mind not having Broadwell quads if and that is IF we have mobile skylake quads in the 3rd or even 4th quarter of this year. I would even expect them to skip Broadwell quads if that were to happen. And this is my thoughts if you are going to bring out Broadwell quads in a couple of months or even a month from now why would I get one when they do become available if Skylake quads will be out in the 3rd or even sometime in the 4th quarter? Again that just my thoughts someone else may feel differently. Now if we don't see any Skylake quads until next year sometime then I would hope to see Broadwell quads within the next month or couple months at least. However its just like we are discussing we don't even know when any Skylake mobile) notebooks will become available to purchase.

 

Have to totally agree with you starlight... BGA is not progressive but regressive IMO... No place for it in high performance notebooks...

 

It seems some people are over-thinking the whole BGA/PGA issue.

The main reason why companies have switched to BGA is because there simply isn't enough market demand.

 

Incorrect. It's the almighty dolla'. Cheaper to manufacture and more profit for the laptop makers. No adapters, less connections, less solder, blah blah blah. It's all about the OEM's more than anything. They don't give a rat about the end consumer.

 

Do you think Clevo could have built P650SE with socketed CPU and GPU and maintain its same thickness? Notebookcheck asks that "why are the CPU and GPU soldered onto the motherboard" in their P651SG review (GTX 980M). So I assume they could have done that but they didn't?

 

It could be argued that less demand for PGA has taken pressure off manufacturers to keep producing PGA parts. Personally, I see the cpu as the last thing I'd upgrade in any laptop i might buy.

That being said, the push these days from most consumers' point of view is for lower priced laptops, not cutting-edge performance; they want to buy the cheapest laptop that'll work, not spend as much as they can budget on a high-performer. A lot of people expect low-priced laptops to perform more than adequately (people call me up at work all the time asking why their £200 laptop isn't as fast as they expected it to be). This of course gets passed onto oems who then do whatever they can to push the price down, which is what we're seeing here.

 

It's not about thickness. I don't disagree that there is a place for it, but honestly if my P650SE was 0.1" thicker I really wouldn't care. I'm holding out to get the P750ZM with the desktop CPU with Skylake and Pascal GPU. Otherwise I'd be on that bandwagon right away.

That's all fine and dandy, and for general purpose laptops that's fine. But for performance machines, for hobbyists and gamers, there is a huge desire/need for socketed components.

 

Intel's IDF Event in Shenzhen China was packed with Little Revelations for the Second half of the Year

Intel showcase wire-free laptops

 

I am not going to repeat myself. I agree with what you're saying, but it is not BGA I am praising.

Please re-read my posts on this issue, I think I was very clear. Having the ability to do something is not better by default (specifically; upgrading cpu+gpu...), especially when all the other benefits are taken into consideration.

 

Whole problem with this "wire-free" technology is if you travel then you still have to take not only a power adapter but a charging pad too. It's OK if you have all day battery power and charge at home at night, but until these "charge pads" become commonplace and integrated with hotel rooms, in airports, even in cars, it won't do a whole lot of good.

 

Enthusiasts/gamers represent only a tiny fraction of overall laptop sales. And even then, plenty of gaming laptop buyers don't know/care about PGA. ('Thin' and 'light' seems to be the new craze among laptop gamers.) The views of NBR users do not represent the whole laptop gamer population.

Thus there is no (significant) market demand for PGA.

You can argue till the cows come home (believe me, I would also rather have PGA), but at the end of the day its the executives at top companies who get to call the shots, as their main objective is to cut costs/maximize profits (which ditching PGA will in my opinion), and make shareholders happy.