Here’s a first look at the battery life of HP’s Ryzen-powered Envy x360

My initial tests of AMD's Ryzen 5 2500U APU gave us a fine picture of the APU's performance, but we admittedly didn't test battery life in that initial article. Part of the reason for that omission was to avoid drawing unfair comparisons between the 15.6" HP Envy x360 that plays host to the Ryzen 5 2500U and the 14" Acer Swift 3 machines we used to represent Intel's Core i5-8250U.

Although the jump in screen size might not sound large on paper, the practical effects of the Ryzen system's bigger screen on battery life are likely quite significant. Getting the same light output from a bigger panel requires more power, as just one variable, and the HP system also has a pen digitizer in its LCD panel with unclear power-management characteristics.

So long as we keep those caveats in mind, though, we can at least offer a basic picture of how long the Ryzen 5 2500U lets the Envy x360 run on battery using a couple different tests. The one quirk of our test rig compared to the $750 default configuration you'd get off the shelf at Best Buy is the Samsung 960 EVO 500GB SSD we're using as our system drive. A hard-drive-only Envy might run for less time, though Windows' power-management features should generally take the mechanical hard drive out of the equation when it's not in active use.

To run these tests, I set the Envy's screen brightness to 50% and left Windows on its default Balanced power plan. The only changes we made to that default configuration involved disabling the operating system's Battery Saver safeguard and forcing the screen to remain on over the entire course of the test.

First off, I ran the Envy's battery down with TR's Browserbench. This benchmark runs a loop of an older version of our home page with plenty of Flash content sprinkled in, along with some cache-busting code to make for more work on the test system. Browserbench is getting up there in years, but it is repeatable and still offers a decent proxy for light web use. We're working on a new version of Browserbench that runs through a range of real-world web sites, but for now, the old version will have to do.

When the Envy shut off automatically with 5% remaining in its juice pack, it registered six hours and 12 minutes of battery life under Browserbench. That's well short of HP's claimed 11-hour battery life figure, but it would at least get you from New York to Los Angeles on in-flight Wi-Fi. To try and see just how good that figure was among similar machines, at least, I scoured the web for reviews of comparable PCs.

 

Reviews of Envy x360s with Intel eighth-gen processors inside remain scarce, but I did find that an Intel Kaby Lake-powered Envy x360 with a configuration and battery similar to that of our Ryzen system turned in six hours of web-browsing battery life for the folks at Laptop Mag. That test suggests Ryzen Mobile could be delivering competitive battery life against similar Intel systems, but it's hard to say just how competitive the Ryzen 5 2500U is without more in-depth (and possibly less representative) directed testing on our part.

Web browsing isn't the only use of a machine on the go, of course. To test video playback, I set up Windows 10's Movies and TV app to loop the 1920×1080, 55 Mbps H.264 version of the Jellyfish reference video until the Envy's battery died. I confirmed that Movies and TV was firing up the GPU's video decode engine using the GPU-monitoring tools in the latest version of Windows 10's Task Manager before letting the test run. Incidentally, here's a full accounting of the Ryzen APU's video-decoding capabilities, as ferreted out by DXVA Checker:

After displaying four hours and 37 minutes of pulsating sea life, the Envy x360 went dark once more. That result parallels the four-hour-and-32-minute run time achieved by the folks at HotHardware in their video playback testing, although the site claimed it had to run its x360's screen at 100% brightness to achieve a comparable output level with its other laptops. While I didn't run our Envy x360 so brightly, HotHardware's results still suggest we can be confident that our run time is in the right ballpark.

With these two tests in the bag, it seems like our 15.6" Ryzen system delivers only average battery life for its size class. I'd still caution against drawing too many comparisons between the Envy x360's battery life and that of other laptops at this stage, though. Implementation differences matter, and we don't know how Ryzen Mobile will behave in smaller and lighter systems. It's still early days for drivers and firmware, too. Our preliminary results and research suggest that Ryzen Mobile's battery life could be competitive with that of similar Intel systems', though, and that's as good a bit of news for AMD as its chip's well-rounded performance.

Comments closed
    • wownwow
    • 2 years ago

    Below is the battery test that both are HP Envy x360:

    [url<]https://www.laptopmag.com/articles/amd-ryzen-mobile-vs-intel-8th-gen-core[/url<] 5:17 for Intel-based and 5:11 for AMD-based.

    • Unknown-Error
    • 2 years ago

    So let me get this straight? This stupid AMD GPU does not know when or how to power down? Doesn’t the GPU have any power or clock gating?

      • wow&wow
      • 2 years ago

      It’s the GPU for dummies or so smart that only Raja Koduri knows how to use it 🙂

      Maybe HP doesn’t know how to design for low power consumption or having learned from Dell CEO HP CEO has a secret deal with Intel CEO 🙂

    • Rza79
    • 2 years ago

    Questions:
    I assume you’re using Win10 v1709, which includes the Ryzen power profile. Does this HP laptop use that profile or just the balanced power profile?
    I’ve read that the display supports Freesync after a BIOS update and a reinstall. Does your machine show that Freesync is enabled?
    Too bad that there’s no way at the moment to see at what frequency the gpu runs during video playback.

    • tipoo
    • 2 years ago

    I’ve seen in a number of reviews that 100% brightness on most laptops and tablets doesn’t really hurt end battery life more than a few percent, kind of funny since most people I’ve seen tend to go to barely visible eye squinting brightness when low on battery.

    • hashhakaj
    • 2 years ago

    Thanks for this useful article
    For me, what do you expect from 15 inch laptop. Any laptop with this size can’t hit more than 5 hours unless if it equipped with larger battery capacity.
    Even if this laptop with Intel CPU it will get the same working time on battery, unless the ULV models as, atom, core m and core Ix Y models.

    • Pancake
    • 2 years ago

    Seems my 5.5hr gut-feeling Browserbench guess is right on the mark! I guess this is just your average run-of-the-mill mediocre midrange overly large laptop. The only “Envy” about this laptop is the feeling of the owner when they see superior machines.

    However, people don’t buy laptop processors. They buy laptops. Let’s see if Ryzen Mobile can make it into a desirable laptop. The performance is certainly good enough. Here’s hoping to seeing that combined with respectable battery life.

    • derFunkenstein
    • 2 years ago

    Now that I’ve had time to ponder all this, I’ll post more than a one-word reaction:

    1.) The machine fell WAY short of HP’s claimed battery life.

    2.) Referring back to the i5-8250U review, the dGPU is off for all these tests, right? Optimus would be using the iGPU for web browsing and video playback, I’d hope, so the one place where Ryzen could “shine” (relatively speaking) compared to the MX150 is in gaming. And that seems like a pretty uncommon scenario, IMO.

      • dragontamer5788
      • 2 years ago

      [quote<]so the one place where Ryzen could "shine" (relatively speaking) compared to the MX150 is in gaming. And that seems like a pretty uncommon scenario, IMO.[/quote<] I dunno. If I'm taking a long flight somewhere, I'm more likely to be playing Cuphead or Touhou than I am to be watching a movie. Of course, the last few airplanes I've been on have 120V chargers + WiFi on them, so its not like I really need a long battery life anyway. Since it is a 2-in-1, maybe media playback really is a good and practical test. (Fold the laptop back into tablet mode, then "consume" media on my bed or couch).

        • derFunkenstein
        • 2 years ago

        It is but in any laptop with an Intel CPU i expect it’s handled by the integrated graphics, so the Nvidia chip isn’t a penalty

          • dragosmp
          • 2 years ago

          But it does show the Radeon IGP is less efficient than the Intel IGP. It follows then the Radeon IGP can’t really power down as well as an Intel IGP. The Nvidia mobile GPUs don’t seem to be able to power down as well as an Intel IGP, but that’s no excuse for AMD – Nvidia GPU don’t need to, as they’ll always fallback on an Intel IGP. IMHO the Radeon IGPs should be designed to be able to power down most of the GPU cores and ancilliaries, like core parking/cache shutdown and flush.

          Maybe we’re reading too much into this single datapoint, there are reasons to believe the HP consumes maybe more: a fast spinning fan 1W, more screen 2W, good speakers and amp 1W, and these are always on – so over 6h as much as 24Wh. Maybe the Radeon iGPU is powering down to similar levels as the Intel IGP or more to compensate for HP’s bad component choice.

          In way too many words I’m basically saying that we’ll need a more apples to apples comparison from all POW, not only screen size.

            • semitope
            • 2 years ago

            people always forget the meaning of efficiency when it comes to AMD. its not such a simple thing. The Radeon iGPU is also significantly faster. Efficiency is a measure of performance vs power consumption.

            A caveat is what kind of performance? eg. Radeon GPUs are much better than Nvidia GPUs in certain tasks, so they are more efficient when you also take into consideration the power they use. But people only compare performance in gaming and usually dx11 gaming.

            Efficiency is more complicated than people pretend and it seems you are only considering one factor in your comment. i.e. simply power consumption.

            The differences you are talking about could also be down to the laptop design since they aren’t that big.

            • Airmantharp
            • 2 years ago

            You’re missing something- if it eats too much juice when performance [i<][b<]isn't needed[/b<][/i<], that's bad for mobile. This is a primary concern for Ryzen Mobile, more so than absolute performance, especially for the ultrabook class products. It's also a concern for the quad-core Intel parts, though less so, as Intel has less to optimize.

            • hashhakaj
            • 2 years ago

            Correct. Unfortunately, most people here don’t understand the worth of APU in power efficient.
            How. The test was in 2 things: browsing and video playing. Theses 2 things are not perfect for APU, overall. If you compared it with Intel, the power of APU is worse.

            But, what about playing games? This thing is the most not interested by most people. Ryzen 5 mobile is 25w when playing GTA V 1080P 30FPS while Intel i5 is 15w when playing GTA V 1080P 15FPS. For increase the gaming performance in Intel side, it will equipped with Nvidia low end dGPU 150mx. So the gaming performance will becomes better than APU however, the power is not. The power total for Intel+Nvidia will become 15+25=40w vs. 25w for Ryzen 5 mobile. So, when playing games on battery, who will working longer?

            • derFunkenstein
            • 2 years ago

            If you’re reading anything at all into this comparison your reading too much. Different display sizes, different battery capacities, and so on mean that comparisons are invalid beyond these two machines.

            • cygnus1
            • 2 years ago

            I’ll add another data point. My HP 13″ Spectre X360 (with only an i7 7500U, no dGPU) is supposed to get 10+ hours of battery life. It does not. Get’s maybe about 6 or 7 hours of general use with chrome and ssh type sessions.

            • dragontamer5788
            • 2 years ago

            Seems like HP 2-in-1 Systems might be wasting power somewhere, somehow.

            I’m going to guess Touchscreen + Digitizer.

            • cygnus1
            • 2 years ago

            Yeah, somewhere. Could be the Bang & Olufsen sound system, who knows. It’s definitely less than HP advertises though.

    • Bumper
    • 2 years ago

    wow great turnaround! I really appreciate you testing these things.

    Basically confirms what I thought. I am now considering going with an Intel/NVidia solution.. The battery life would have to be better than the competition for me to consider an apu. Otherwise it’d have to be at least a couple hundred dollars cheaper.

      • derFunkenstein
      • 2 years ago

      Wouldn’t the MX150 be basically disabled since the iGPU could do everything in these tests? The only place I might expect a big difference is gaming tests with the dGPU enabled, but even that’s not a common scenario.

        • Bumper
        • 2 years ago

        idk. Maybe some more testing will give more answers. It needs to have >=intel browsing battery life though. It should have at least an hour more gaming battery life than mx150. If it has those two things it might be worth considering for a hundred dollars less than the i5/mx150. In most things it is slower than the intel part. not-gaming battery life needs to be really close from my perspective.

        I’m kind of interested in the 2700u. if it can approach mx150 gaming performance for slightly less money and get better gaming battery life, I’d strongly consider it over a solution with a more efficient intel part. I wonder how hot the dgpu laptops get in comparison to the apu laptops…

          • derFunkenstein
          • 2 years ago

          You have to normalize for differing battery capacities, display sizes, backlights, etc. It’s not an easy question.

            • Bumper
            • 2 years ago

            Yeah. This hp laptop has a shoddy display anyways. If I have the info sometime by January I’ll be able to make a*good* buying decision.

    • chuckula
    • 2 years ago

    If you are really concerned about the display you could turn it off entirely to do a differential analysis that tests the impact of the display on the battery life.

    Displays certainly do pull a decent amount of power, so taking it out of the equation could be helpful.

      • Jeff Kampman
      • 2 years ago

      Now that we have real-world benchmarks I’ll probably pursue this more academic line of testing.

      • Smeghead
      • 2 years ago

      That’s a pretty decent idea. One tweak to that might be to hook the machine up to an external monitor of equivalent resolution, then disable the laptop’s display.

      With the display off by itself, there may be all sorts of power-saving fun and games that might not deliver representative results. Forcing the GPU to still drive a similar number of pixels, the workload should be fairly equivalent.

      That does assume that the power required to run a displayport/hdmi/whatever link is pretty minimal, however.

      • willmore
      • 2 years ago

      Do you have the ability to just hook a bench power supply to the thing? One that does power calculations such as ‘usage over time’? Might that be a good way to get actual power useage?

        • dragontamer5788
        • 2 years ago

        That’s not the issue.

        How do you normalize for the different amounts of power that goes into a 15-inch screen vs a 13-inch screen? How do you normalize for the active-digitizer and touch-screen, which sucks up power on the x360 Envy but not on other laptops (which are missing those features?)

        Anyone can buy a “Kill-a-watt” meter, pull out the battery, and then measure the power that goes into a laptop. But that’s still not going to make a fair test.

          • willmore
          • 2 years ago

          As has been mentioned several times, turn the display off and plug a reference resolution external display in. Take the whole darn display out of the equation if that’s your concern.

          Also, a Kill-a-watt is a bad idea as it leaves the power brick in the equation and those aren’t all the same.

            • Jeff Kampman
            • 2 years ago

            I’m running the systems through our battery benchmarks on external displays right now. Should be interesting.

            • Mr Bill
            • 2 years ago

            Ah! Work the GPU but not the screen, great idea; and you can even compare for equivalent “resolutions”.

            • willmore
            • 2 years ago

            See, right here, this is why I love TR. They *listen* and they adapt. They use the community to improve.

            Thank you, Jeff! You continue to impress.

      • Mr Bill
      • 2 years ago

      Begs the question, do any of these portables have LED backlighting and would that save a significant amount of power?

        • psuedonymous
        • 2 years ago

        I don’t think even the cheapest bargain-bin laptop has used CCFL backlighting for the last 5 years at least. If nothing else, it’d cost more than LED edgelighting for no benefit.

    • chuckula
    • 2 years ago

    It’s not a pure apples-to-apples comparison so it’s not the only worthwhile data point, but one of the Swift 3 notebooks that TR reviewed a few days ago delivered 9.5 hours of battery life in the browser bench mark. That’s the one that had the discrete Nvidia MX-150 GPU. The i5-8250U by itself did 10.4 hours.

    [url<]https://techreport.com/review_full/32863/intel-core-i5-8250u-cpu-reviewed[/url<] I'd like to see how the different Swift 3 models handle the video loop too.

      • DragonDaddyBear
      • 2 years ago

      For non-subscribers.

      [url<]https://techreport.com/review/32863/intel-core-i5-8250u-cpu-reviewed[/url<]

      • Goty
      • 2 years ago

      There are far too many variables involved in this sort of comparison to try and draw any meaningful conclusions; I think even comparing TR’s results to results found elsewhere for Intel-based Envy x360s are not likely telling of Raven Ridge’s power consumption. The only meaningful conclusion that can be drawn here is that [i<]this particular implementation[/i<] of a Raven Ridge laptop exhibits very average battery life and only further examination of other laptops will reveal what the APU is (or isn't) capable of.

      • Goty
      • 2 years ago

      For those that are interesting in something a little more concrete, [url=https://www.notebookcheck.net/Our-first-Ryzen-5-2500U-benchmarks-are-in-and-Intel-has-every-reason-to-worry.266618.0.html<]Notebook Check[/url<] has actual power consumption numbers for a number of systems, in which the Ryzen 5 2500U posts better numbers in every scenario than its Intel competition (all 15 inch HP Spectre x360s) with the exception of the 15W Core i5-6200U in the 15 inch Spectre x360 15-ap012dx. It should be noted, however, that all of the Intel systems are saddled with higher resolution screens, which I'm sure is inflating their numbers by quite a bit.

        • Klimax
        • 2 years ago

        No wonder Ryzen posts better numbers when it doesn’t have to drive high res display that eats quite bit more watts as well.

          • Shobai
          • 2 years ago

          If this machine had to drive the same display as the others, it’d be a more useful comparison – have we thrashed that horse enough, now?

          • Goty
          • 2 years ago

          The difference shown (around 5W at idle) is about the difference in power consumption I would expect between a typical FHD display and an IGZO 4K display of the same size. If that’s the difference introduced by the display, then we still have to explain away the remaining 15W or so difference between the 8550U and 2500U in the average load power consumption numbers. Some of that is down to the 150MX for sure, but certainly not all of it. I’m not sure if the HDD in the Ryzen laptop or the SSD in the Intel laptop consumes more power, but I expect it’s not a large difference either way.

          • lampuiho
          • 2 years ago

          It seems people failed to realise that most of the other laptops have discrete graphics card. The only one that doesn’t is HP Spectre x360 15.
          I still haven’t found an article that is completely dedicated to investigating the efficiency of either processors.
          Due to differences in implementation of various components, performance per watt needs to be looked at under different work load. Including integer operations (most involved in web browsing), floating point operations (video encoding), single and multi threaded gaming.
          Moreover, there is also a performance per watt curve sweetspot that the TDP is not hitting. Sometimes the CPU is overclocked and go over it while some are underclocked and not hitting that yet.
          For the laptop design team though, I think they only look at TDP but usually the CPUs don’t adhere very closely to the TDP and consume more power than it should because there is sufficient cooling for it to dissipate more heat.

      • Pancake
      • 2 years ago

      Aha! Found a piece of the puzzle. The Swift 3 has a 48Wh battery. The Envy 360 has a 55Wh unit. Now, the 15.6″ screen of the Envy is 22% bigger than the Swift’s 14.1″.

      So, assuming same power per screen area and all the difference in power consumption is for driving the screen then a hypothetical 14.1″ Envy with a 48Wh battery would last:

      6.2 x (48 / 55) x (1.22) = 6.6 hrs

      This is of course an overestimate as obviously a chunk of consumption is for processing and the Envy has a noted dim screen. But let’s be even more generous and round it up to 7 hours.

      So, the least efficient i5-8250 with MX-150 combo is

      9.5 / 7 = 1.35 = will get 35% more mileage per tank of lithium-ion gassy goodness.

      Result = pure massacre. These are both likely lightly modified reference designs from AMD/Intel. Don’t expect thin AND light AND long lasting Ryzen mobile laptops anytime soon. The massive difference in power efficiency will mean Intel gets into the premium laptops. AMD will languish in the el-cheapo category.

    • derFunkenstein
    • 2 years ago

    Yikes

    • chuckula
    • 2 years ago

    So much for throwing VLC under the bus as the cause of the poor battery life when even the default Windows 10 video player is showing the same behavior.

      • Anonymous Coward
      • 2 years ago

      You know, even if it does turn out that Raven Ridge is inefficient at decoding video, this emotionally-driven finger-pointing will [i<]still[/i<] not be worthy of any respect.

        • Airmantharp
        • 2 years ago

        If it’s inefficient at video decoding, it’s simply not competitive. That’s a very basic thing that needs to be gotten right.

        • IGTrading
        • 2 years ago

        I don’t think VLC is to blame.

        It might be something within the driver or firmware or even VLC, but since it’s software, it will soon be fixed.

        So AMD’s showing will be even more impressive as problems get ironed out.

        We need to keep in mind that this is the absolute first VEGA Mobile chip and driver and nVIDIA had more than 3 years to tune its mobile solutions.

        I think that AMD’s drivers will show a serious improvement in the following months, just like the usually do.

        If you think of it, it’s the “tradition” of the company 🙂 just like AMD Radeon Fury X which was almost always a few steps behind GF 980 Ti when launched, but today it beats the GF 980 TI in 99% of all games at both FullHD and 4k UHD.

        With Raven Ridge, being such a complex solution and having to show both performance and frugal power consumption at once, maturing drivers will show just how powerful it truly is.

        Plus you can’t really blame AMD for not having even better battery power when they show almost double the 3D performance of ALL Intel iGPUs.

        It would be ridiculous to expect 200% the performance @ lower power consumption @ lower price, but AMD will likely get way closer to that goal.

          • dragontamer5788
          • 2 years ago

          [quote<]since it's software, it will soon be fixed.[/quote<] This is overly optimistic and I'm incredibly against this point of view. Case in point: Overwatch (one of the biggest titles in recent history) has a gamebreaking driver bug [b<]right now[/b<] which causes the game to crash on AMD systems: [url<]https://us.battle.net/forums/en/overwatch/topic/20758607000[/url<] Some software issues will take higher priority than others. In the great scheme of things, "uses slightly more battery than expected" is rather low compared to keeping the drivers stable on major AAA video games. I'm 100% sure that this Overwatch bug will be fixed, due to its high visibility. But that means AMD's software resources are strained on this [b<]other[/b<] bug, not working on the video driver battery issue. [quote<] I think that AMD's drivers will show a serious improvement in the following months, just like the usually do.[/quote<] And eventually, the next generation (ie: Navi) will come out, and AMD's software resources will be spent on that instead. The majority of improvements will be seen in common AAA-titles with high-visibility... and not on very specific issues. You buy the system as it is functioning, right now. Its fine to hope for the best, but its not good to make a purchase decision based on hopes and dreams alone. Sometimes, software issues are never fixed, even for years. With that said, we don't know AMD's internal process for prioritizing bug reports and all that. But there are always a torrent of bugs and minor issues associated with software and/or drivers. Not all of them will be fixed by next year. So its best not to hope that things get fixed IMO, but rather... its better to make decisions based on how the system works today (software bugs and all).

            • IGTrading
            • 2 years ago

            Like AMD has shown with Radeon Fury X the performance has improved continuously despite the launch of Polaris and Vega.

            A video decoding battery bug would affect all AMD’s mobile platforms, so I think it is even more visible then the all-important Overwatch 🙂

            We’ll see.

      • brucethemoose
      • 2 years ago

      Who was claiming that? VLC is DXVA now, AFAIK.

    • willmore
    • 2 years ago

    Sadly, I don’t see AV1 in there. I hope we see more of that in the future.

      • Anton Kochubey
      • 2 years ago

      The bitstream format isn’t even finalized yet – don’t expect hardware decoders until at least late 2018 (or maybe even early 2019).

      • brucethemoose
      • 2 years ago

      AV1 would just be a CPU test.

      Besides, have you even [i<]tried[/i<] encoding it yourself? It's sooooooo slow right now.

    • dragontamer5788
    • 2 years ago

    Good review, good tests, good research.

    DXVA Checker was a great touch after you’ve confirmed the HotHardware video test.

    • Unknown-Error
    • 2 years ago

    Watching sea life for that long?

      • willmore
      • 2 years ago

      Why, you jelly?

Pin It on Pinterest

Share This