Intel promises 50% battery life gain for Haswell laptops

Haswell is coming out early next month, and it’s everywhere in the news. Among other things, motherboard makers have been tripping over themselves to show off their upcoming Haswell-ready mobos. We’ve heard comparatively little about the mobile incarnations of the chip, but in a briefing yesterday, Intel gave the guys at PC World a few nuggets of info about upcoming Haswell laptops.

According to the site, Intel said those notebooks will bring about a whopping 50% battery life increase over current, Ivy Bridge-based systems. In idle or standby mode, run times will purportedly increase 20-fold. And even more impressively, Intel claims those gains won’t involve any performance tradeoffs.

The chipmaker didn’t mention which classes of laptops will see those gains. Considering how hard the chipmaker has been pushing ultrabooks, though, I think it’s safe to expect the improvements there. We already know that some mobile Haswell variants will have thermal envelopes as low as 10W, down from 17W for current ultrabook-bound offerings.

For what it’s worth, we clocked Asus’ Zenbook Prime UX31A, a premium ultrabook, at 7.5 hours in our web-browsing battery life test. A 50% gain over that would work out to 11.3 hours. Even if Intel is quoting a best-case scenario and real-world gains turn out to be more modest, we may still end up dangerously close to the 10-hour mark. That wouldn’t be bad for a three-pound system with a 13", 1080p display—not bad at all.

Comments closed
    • Krogoth
    • 6 years ago

    50% is a wee-bit optimistic. The real world will probably be closer to 20-25%. It is not too bad for portable platforms. This is where the new power states and integrated voltage regulator matter.

    Haswell is a die-hard portable CPU at heart. The desktop version is going have a decent amount of OC’ing headroom. You could simply throw a huge HSF on it with near-silent fan at stock clockspeed if silence is your thing.

    • albert 89
    • 6 years ago

    I’d like laptop batteries to stay at pre-power efficiency levels [2012?] as they’ll be at maximum power through-put that way they get recharge at say once a week !!!!!!
    Now that would be nice ! If batteries were to shrink in line with power efficiency delivered by the latest CPU’s & APU’s for the sake of aesthetics then you’d end up recharging the bitch every 3 hrs like before all because your unit is now 1mm square smaller ???? Which defeats the term ‘mobility’ !
    I mean tie me up, pull the pin put a grenade in my mouth stuff me in a barrel and roll me down the street for fuck sake.

    • drsauced
    • 6 years ago

    Uh, hm. Marketing fluff. By far the biggest power drain on a laptop, phone, tablet, is the screen. A %50 cut in battery drain from the screen would be something to write home about, but a %50 cut in CPU power is pretty much a snooze.

    Beep. Beep. Beep. *SMACK!

    • DavidC1
    • 6 years ago

    abw:

    Typical PC Desktop enthusiast that NEVER reads about anything else, or is interested about it says this.

    It’s true, for Desktop, and even for big Notebooks with quad cores and Discrete graphics, Haswell doesn’t look much.

    But that’s not the point of the chip. Intel is far away from days of sacrificing everything for little bit of performance, because the sacrifice in portability and battery life is not worth it for the majority.

      • abw
      • 6 years ago

      I got only laptops since 2001 , mind you…..
      Amazing how people can draw so ill conceived assumptions…

      Celeron 600M
      Duron 800M
      Athlon XPM 1800+
      Pentium 4M Northwood 1.6G
      Pentium T4400

      Currently looking for a Netbook with 11.6 screen ,
      Kabini 4C/15W should fit the purpose.

    • ssidbroadcast
    • 6 years ago

    This is good news. My HP Spectre XT runs an Ivy Bridge Core i7-3540U and the battery life is pretty miserable. Probably 2 hours… you [i<]might[/i<] squeeze three if you turn the brightness all the way down and keep the volume muted.

    • esterhasz
    • 6 years ago

    Interestingly though, the efficiency gains seem not to translate to the desktop version. Heise (some of the best technology reporting there is) went out and bought a i7-4770 in a store (yeah, well) and ran some tests: [url<]http://www.heise.de/ct/meldung/Fruehstart-Intels-Haswell-im-Kurztest-1868226.html[/url<] Higher system power than Ivy at load; AVX2 flies, though. Hopefully at least idle power will be down...

      • nico1982
      • 6 years ago

      Well, it is rather obvious that the 50% improvement is achieved by specific parts under specific workloads and not equally spread across the entire lineup under any circumstance. Getting an ULV chip going from 8 to 12 hours of playback is reasonable, achieving the same on a top of gamma 4C/8T part under cinebench is going to be another matter altogether. Other combinations of silicon/scenario will bring even different results.

      • Voldenuit
      • 6 years ago

      I’d imagine that some of the power saving states on the desktop are disabled to maintain compatibility with motherboards/PSUs. On top of that, the standard BIOS/UEFI configuration on most consumer motherboards would not be as tuned to power saving as those on laptops. Lastly, it’s very likely that intel is applying different binning requirements for laptop vs desktop chips.

      So no surprise that we are not seeing the same efficiency gains on the desktop.

      • DavidC1
      • 6 years ago

      The reason is because only the ULT, or otherwise known as Ultrabook-bound parts, get all the power savings.

      Desktops and regular Notebooks will have regular, incremental gains.

    • Chrispy_
    • 6 years ago

    50% gain is quite a claim, considering that in a typical laptop things like the screen, memory and storage consume close to half the available power, regardless of what the CPU is doing.

    The lower the TDP of the processor, the more important it is to shave power elsewhere. My smartphone often tells me that since I unplugged the charger, more than two-thirds of the battery has been spent on the screen alone.

    The good news is that Intel’s IGP actually looks viable for use now, but

      • axeman
      • 6 years ago

      Seconded. When I read the headline I thought, did Intel come up with some sort of breakthrough display tech?

      • DPete27
      • 6 years ago

      Very true. But since displays are slow to improve efficiency, we’ll have to settle with CPU/GPU efficiency improvements for now.

      This also makes me wonder: 17W -> 10W = %41 CPU power reduction. (I know there are deeper factors than straight TDP, such as IPC improvements and faster-to-idle that come into play here)

      Let’s assume that they’ve managed to eek out an extra 1.5W power savings elsewhere (maybe power delivery or chipset) to make it an even 50%. Does that mean Intel is making their 50% improvement claims by testing with the display off? Tricky tricky.

      …Then we get to [i<]Shambles'[/i<] comment about laptop makers using the power efficiency improvements to save themselves money with smaller batteries...

    • flip-mode
    • 6 years ago

    [quote<]In idle or standby mode, run times will purportedly increase 20-fold. [/quote<] Just to make sure I'm reading that correctly, let's take a hypothetical 8-hour idle run time, times 20, equals a 160-hour idle run time? Is that the correct interpretation?

      • peartart
      • 6 years ago

      That’s probably only for the CPU, so antennas and any other power draw would pull it down.

      • chuckula
      • 6 years ago

      If you are turning off the rest of the components in the notebook, that should be possible. If you are still running the display/wifi/etc., then Amdahl’s law comes into effect. As we have already seen in smartphones and tablets, even a theoretically impossible zero-power CPU can only improve the battery life to a limited extent when there are other power-sucking components attached to the battery.

        • MadManOriginal
        • 6 years ago

        I thought Amdahl’s law referred to non-linear improvement in multithreaded applications when using more cores, so 2x the cores doesn’t mean 2x the performance. But otherwise, yeah..

          • chuckula
          • 6 years ago

          Amdahl’s law is really just a restatement of the law of diminishing returns that is commonly applied to lots of issues in CS. The original context was in the area of parallelized software, but it applies to many aspects of computer hardware & software.

            • peartart
            • 6 years ago

            Alternatively, it’s a variation of the mathematical fact that the sum of non-negative numbers is zero if and only if all the numbers are zero.

      • NeelyCam
      • 6 years ago

      Actually, WSJ (where I first read about this) said that at idle it’s 2-3x better:

      [url<]http://blogs.wsj.com/digits/2013/05/23/intel-discloses-power-savings-of-next-chip-line/[/url<] But even then, remembering that "AMD defines 'all-day' battery life as 8+ hours of battery life in Windows idle" makes Haswell sound incredibly good.

      • xeridea
      • 6 years ago

      A 20 fold increase in standby life is impossible with current technology. In standby, the main power draw is the RAM because it requires some power to keep it alive. With some memory technologies in development may be, but the CPU is already basically consuming 0 power during standby anyway, this has been the case ever since sleep/standby mode on computers was invented.

        • NeelyCam
        • 6 years ago

        It’s not 20x. PC World got it wrong.

        The [url=http://blogs.wsj.com/digits/2013/05/23/intel-discloses-power-savings-of-next-chip-line/<]Wall Street Journal blogpost[/url<] said it is 2-3x better compared to Ivy Bridge, and 20x compared to Sandy Bridge. Although, the article seems to use 'idle' and 'standby' interchangeably... so I'm not sure if the journalist understood everything correctly. [url=http://www.eetimes.com/electronics-news/4415059/Intel-Haswell-packs-integrated-voltage-regulator<]EE Times[/url<] also quotes 2-3x vs. Ivy Bridge: [quote<]"Over the past year, Haswell chip design teams worked with Intel’s fab teams to optimize the version of its 22nm process, called P1270, used for the x86 chips. Intel would not describe the enhancements, but it said their effect was to lower transistor leakage by 2-3x compared to its previous Ivy Bridge generation while lowering its minimum voltage level (Vmin) and not reducing the chip’s frequency."[/quote<] If this is the actual statement, then it doesn't necessarily mean that idle power is 2-3x lower on Haswell vs. Ivy Bridge... I have a feeling that a non-technical journalist didn't understand it right, wrote an incorrect article, and then it got propagated across the interweb. However, the 50% battery life improvement under load seems like an accurate statement, and "lowering its minimum voltage level (Vmin) and not reducing the chip's frequency." seems to confirm that the improvement really is coming from the lower supply voltage as I suspected

          • abw
          • 6 years ago

          Chasing the usual fairy tales..??..

            • NeelyCam
            • 6 years ago

            Huh.,.?

            • abw
            • 6 years ago

            huH.,.!

            • chuckula
            • 6 years ago

            I HAVE LEARNED ABW’S LANGUAGE OF INTERPOSED PERIODS AND COMMAS!

            [b<].,,.,..,.,.,,.,.,.,.,.,.,.,....,..,,,,.,.,,.,.,.,.,.,...,.,.,.,.,.,.,....,,,.,.,.,[/b<] That was the prologue to War & Peace!

            • abw
            • 6 years ago

            Still desesperatly lacking substance and deepness….to the point
            that you missed the essential of this language but wasnt it expected ,
            indeed…?…

            • MadManOriginal
            • 6 years ago

            Why do all your posts look like bad attempts at poetry?

            • NeelyCam
            • 6 years ago

            It’s the Miller [b<]high[/b<] life

            • abw
            • 6 years ago

            It s unwillingly and just an impression , actualy i m rather
            lazy when it comes to type , hence the broken style…

            • chuckula
            • 6 years ago

            Hey ABW… here’s a video of a “fairy tale” running continuously for 9.5 hours: [url<]http://www.youtube.com/watch?feature=player_embedded&v=Rs7sWZQRhic[/url<] So in whatever language you speak (I'm guessing North Korean), "fairy tale" translates to "shipping Intel hardware" right?

        • WillBach
        • 6 years ago

        It’s not impossible. It’s already present in Atom Windows 8 tablets (and in iOS and Android tablet, too). You turn the screen off, put it down, pick it up days later and still have a charge, with your email up-to-date and everything.

      • WillBach
      • 6 years ago

      That’s talking about the system-level power during sleep and so called connected idle. If you turn an iPad’s screen off and put it down for a 20 hours, it still has most of its battery left when you pick it back up. The Atom tablets on the market support that with Windows 8, Intel is bringing that feature set to the big cores with Haswell.

    • Shambles
    • 6 years ago

    So Basically we can expect 50% smaller batteries in the never ending fools quest to have thinner and lighter laptops at the expense of practicality?

      • MadManOriginal
      • 6 years ago

      Batteries 1/3 smaller would make it equal to today’s run times….but yeah, that’s one of the first things I thought of too. Laptop makers have made us all too cynical.

        • Shambles
        • 6 years ago

        Ugh, butchering math already this morning. Bad way to start a day :S

          • MadManOriginal
          • 6 years ago

          No worries, I +1’d your comment because I can totally see that happening.

      • Chrispy_
      • 6 years ago

      I really hope you’re wrong, but the cynic in me is correct far too often in matters like these.

      • Aliasundercover
      • 6 years ago

      50% smaller and glued in deep where they can’t be swapped.

      • slowriot
      • 6 years ago

      Pretty much my first thoughts as well… the current state of laptops makes me sad. It seems nearly every laptop has lost functionality for the sake of marketing check boxes.

      • HisDivineOrder
      • 6 years ago

      So, so true.

      • ptsant
      • 6 years ago

      It’s not over until your laptop can be classified as a cutting instrument and be carried by a 3mo baby. In fact, you’ll notice that the majority of 13.3″ notebooks are now constrained by the horrible “ultrabook” factor and do not provide such luxuries as upgradeable RAM or dedicated graphics.

        • cynan
        • 6 years ago

        Being banned from carry-on due to being a potential bladed implement of terrorism might be the only thing that stops this insane spiral of mobile devices into anorexic madness.

        • confusedpenguin
        • 6 years ago

        Probably easier just to hand a baby a knife.

      • gmskking
      • 6 years ago

      Why improve the batteries when they can sell you replacements and backups instead. The fundamental flaw of a capitalistic society. We are witnessing the fall of it. I for one am glad.

    • Unknown-Error
    • 6 years ago

    I am 100% about buying a Haswell notebook. The Question is when is it going to available for the masses? I’ve been waiting a long time for this.

    I am thinking of either i7-4800MQ or i7-4900MQ with a discrete card. I don’t think I’ll go for GT3e i7-4850HQ or i7-4950HQ.

    Any advice people?

      • Sargent Duck
      • 6 years ago

      Wait till benchmarks are released?

      • Unknown-Error
      • 6 years ago

      someone down voted me for that? That wasn’t even troll comment.

        • moog
        • 6 years ago

        AnandTech is WAY BETTER than TR.

        • NeelyCam
        • 6 years ago

        It’s because you sounded like an Intel fanboi. Any self-respecting techie would either go with Kabini/Richland now, or wait for Kaveri.

      • auxy
      • 6 years ago

      This is really the kind of question that should go into the forums, anyway. Hehe.

    • Chrispy_
    • 6 years ago

    [b<]This comment intentionally left blank[/b<] [i<](reply fail to Shambles' comment, now fixed)[/i<]

      • willmore
      • 6 years ago

      Reply fail?

        • Chrispy_
        • 6 years ago

        Oh yes 🙂

        “Nothing to see here, move along now….”

          • willmore
          • 6 years ago

          Your Jedi power have no effect on me.

    • abw
    • 6 years ago

    Yet another extraordinary claim….from Hasfail to boastwell…

      • chuckula
      • 6 years ago

      Look you moron, It’s [b<]HASBEEN[/b<]! Get with the program you paid Intel shill!

        • abw
        • 6 years ago

        Calm down intel sucker and perhaps you ll get some intelmeds….

      • abw
      • 6 years ago

      I was right , yet another fairy tale for the gullible intel sucker…..

        • chuckula
        • 6 years ago

        [quote<]I was right , yet another fairy tale for the gullible intel sucker.....[/quote<] So since you wrote that in response to your own post, you obviously believe yourself to be a gullible intel sucker. Thanks for playing!

          • abw
          • 6 years ago

          The answer was yet to come but it came , isnt t..?…

            • NeelyCam
            • 6 years ago

            ??

            It must be Miller time wherever you are

            • abw
            • 6 years ago

            Sorry man , i do not live in your imaginary , Miller does mean nothing
            for me , unless you re talking of Miller compensation or theorem….

            • NeelyCam
            • 6 years ago

            You have 6 really bad beers and try to design an opamp, only to realize it’s unstable. Then you do Miller compensation: have 6 more really bad beers and try again

            • abw
            • 6 years ago

            Guess that you knew that i did know that you would
            understand what i was talking about….

    • chuckula
    • 6 years ago

    I really hope that these innovations mean that they can make notebooks thin enough to shave with. You should have seen my disappointment when I got a RAZR.

    • R2P2
    • 6 years ago

    I think the more likely case is that laptop manufacturers will use smaller batteries. That’s how it goes with phones, right? More efficient hardware means a smaller battery and a thinner device, instead of a device that runs longer.

      • tipoo
      • 6 years ago

      I hope not. I never got why care more about 5mm of thickness being shaved off to claim being the worlds thinnest smartphone, rather than keep the old size and use the space for a battery that you may not have to charge every single day, if not more.

    • tipoo
    • 6 years ago

    Did they not say something similar for Ivy Bridge? I’m genuinely asking, I seem to recall something like that. But 50%? Most of the power draw in a laptop is the screen, how can just a processor with lower idle power increase life by 50%?

      • chuckula
      • 6 years ago

      [quote<]Did they not say something similar for Ivy Bridge?[/quote<] They said that Ivy would lower power consumption, and it does. However, they never made promises anywhere even close to 50% savings under load and 2x-3x savings in idle. Of course, you still need to look at the actual battery in the notebook too. Hopefully the OEMs won't further neuter the batteries just because Intel made a more efficient design.

        • peartart
        • 6 years ago

        Oh, they will. Cheaper and lighter is irresistible for OEMs.

      • OneArmedScissor
      • 6 years ago

      Haswell supports Panel Self Refresh, where a memory chip in the panel itself stores the frame buffer.

      That can allow almost the entire computer to effectively go into sleep mode at any moment, no matter how brief, that you aren’t moving the mouse, pressing keys, or refreshing the screen.

      As always, it’s dependent on OEMs to actually implement these things.

        • Star Brood
        • 6 years ago

        I was going to write something about the Panel Self Refresh. I’m glad I’m not the only one.

        • xeridea
        • 6 years ago

        Most of the power consumed by LCDs is the backlight, so this would be a small gain. The CPU is already going to go into a low power mode if nothing is happening.

          • Bubster
          • 6 years ago

          Ivy bridge still idles at a couple of watts.

          • OneArmedScissor
          • 6 years ago

          I don’t know how I could reword that to be any clearer. It is not an issue of the CPU.

      • OneArmedScissor
      • 6 years ago

      Another advantage of Haswell is Intel’s emphasis on firmware tweaks for reducing power of additional controllers. That is finally worth the effort because integration has simplified and standardized so much.

      Intel also designs a lot of the mobile hardware. They surely have new wifi cards for Haswell. Those can be battery guzzlers.

      New, lower power SSDs are also likely, Intel or not. Haswell supports DEVSLP or “device sleep,” which only the newest drives have.

      The Haswell CPU package is not so far off from a hypothetical cube of silicon, but there is still a lot that can be done to make the other parts of a computer into a more dynamic system.

      Think the jump from SpeedStep to turbo states, but for the screen, modem, storage, RAM, and all the interconnects that tie it together.

    • hoboGeek
    • 6 years ago

    I’m impressed! I guessed they focused their attention on lowering the power as opposed to increasing performance as it was till recent years. For a laptop, I believe current calculating power was really good enough, if not more than that. So it was time to put the consumption to a severe dieting program…
    Good job Intel!

    • sonofsanta
    • 6 years ago

    50% is incredible. I can believe the standby gains with the new idle state, but how they’ve pulled off that gain under load is beyond me.

    Are the mobile chips due to be released in June along with the first desktop chips? Last time round the dual-core desktop i3s were delayed until September.

      • NeelyCam
      • 6 years ago

      [quote<]I can believe the standby gains with the new idle state, but how they've pulled off that gain under load is beyond me.[/quote<] I'm thinking it's because of lower supply voltage - when I saw those overclocking results at 0.9V, my immediate reaction was that "this thing is going to have great load power efficiency". Also, the new integrated voltage regulator lets the chip go idle and wake up faster, so the chip can use that to save power in more real use cases (like browsing) than before.

    • hasseb64
    • 6 years ago

    Let’s hope that INTEL deliver this time! Last chance to stop the x86-slaugther.

      • tipoo
      • 6 years ago

      x86 slaughter? You mean the Chromebooks literal 0.02% market share? If you mean the slump in desktop/laptop sales and rise of tablets, they’re expected to rise again next year. And besides, Intel is bringing x86 to the ARM space, so is AMD.

      • Klimax
      • 6 years ago

      You got extra ‘s’ there…

Pin It on Pinterest

Share This