Skip to main content

Snapdragon-Powered Laptops Could Soon Get a Big Boost

(Image credit: Laptop Mag)

It appears that Qualcomm's looking to solve one of the biggest issues that plagued the first Windows laptops that ran on its chips: sluggish performance. New reports suggest that a suped-up Snapdragon 1000 CPU is in the works, which could provide enough speed to compete with systems running Intel's mainstream processors.

The news comes from the Berlin-based WinFuture, which claims that the ARM chips will sport enough performance to be compared to Intel's stronger Y- and U-series x86 processors. The Y-series chips are less powerful and found in Intel's thinner systems, while the U-series processors are used in more-average, everyday laptops.

The WinFuture article also notes that the Snapdragon 1000 will draw 6.5 watts of power, and that it will be a part of a system-on-chip design that draws 12 watts in total. We note this power usage detail because it brings up a very big question about battery life in these Qualcomm laptops.

Intel's U-series chips use 15W of power and run in laptops with 5 to 17 hours of battery life, and its Y-series chips use 4.5W of power and feature in laptops with 5 to 9 hours. Therefore, a 6.5W CPU would make for a laptop with battery life somewhere within those ranges, which could fall below the all-day claims that Qualcomm's made in the past.

For example, the Snapdragon 835-based HP Envy x2 lasted 14 hours and 22 minutes on the Laptop Mag battery test and the Snapdragon 835-based Asus NovaGo made it 12 hours and 19 minutes on that same test.

WinFuture references a newly-developed power management chip to handle the increased needs of this chip and how it consumes energy, so that may be mitigated.

While the Snapdragon 835 laptops also benefited from ubiquitous internet, thanks to LTE connectivity, their speed wasn't the only issue. These machines are still awaiting support for 64-bit apps, meaning that they may not be a great fit for you, depending on the Windows 10 apps you rely on.

This story was originally published on Laptop Mag.

  • mlee 2500
    The conclusions in this article may or may not be right, but they are extrapolated from entirely faulty logic. Namely, when, how long, and how often is that power draw applied.
    Reply
  • alextheblue
    21088288 said:
    The conclusions in this article may or may not be right, but they are extrapolated from entirely faulty logic. Namely, when, how long, and how often is that power draw applied.

    Yeah. For all we know the Snapdragons have better (more granular, faster) power/clock adjustments. They also might be more efficient at the majority of the power curve, merely having a similar top-end draw. It also doesn't factor in performance. A higher-performance chip won't have to spool up as much and/or as long to complete the same task, so if draw is the same but performance is higher you can sleep more. Similarly, the number of cores and core configuration are big factors. If it's big.LITTLE there will be opportunities to shift low-demand workloads to the more efficient small cores. Speaking of which, workload is always a factor. Basically just looking at TDP ranges isn't enough to guess battery life.

    Even the SD850 will be a pretty substantial upgrade over the 835 devices. With that being said, the biggest factor for ARM in terms of raw performance is how much software (be it UWP or Win32) can they get native compiles for.
    Reply
  • electronicbineries
    ARM CPU should stay Android even for desktops/Tablets/Notebooks.

    The App store alone is enough reason.
    Reply
  • mlee 2500
    21088841 said:
    21088288 said:
    The conclusions in this article may or may not be right, but they are extrapolated from entirely faulty logic. Namely, when, how long, and how often is that power draw applied.

    Yeah. For all we know the Snapdragons have better (more granular, faster) power/clock adjustments. They also might be more efficient at the majority of the power curve, merely having a similar top-end draw. It also doesn't factor in performance. A higher-performance chip won't have to spool up as much and/or as long to complete the same task, so if draw is the same but performance is higher you can sleep more. Similarly, the number of cores and core configuration are big factors. If it's big.LITTLE there will be opportunities to shift low-demand workloads to the more efficient small cores. Speaking of which, workload is always a factor. Basically just looking at TDP ranges isn't enough to guess battery life.

    Even the SD850 will be a pretty substantial upgrade over the 835 devices. With that being said, the biggest factor for ARM in terms of raw performance is how much software (be it UWP or Win32) can they get native compiles for.

    Compiled and OPTIMIZED. I cannot speak for the Windows code base, obviously, but I do know that with the Linux and OpenSource ports to Qualcomm's Centriq line of Server chips it wasn't enough for us to just re-compile the code. Most compilers and source code have decades of optimization geared towards x86 architectures, putting entirely new platforms at a bit of a disadvantage.

    Regarding the power draw, Qualcomm has an outstanding track record of disabling parts of the SOC that aren't needed...perhaps that will give them an edge.
    Reply
  • alextheblue
    21092285 said:
    21088841 said:
    Even the SD850 will be a pretty substantial upgrade over the 835 devices. With that being said, the biggest factor for ARM in terms of raw performance is how much software (be it UWP or Win32) can they get native compiles for.

    Compiled and OPTIMIZED. I cannot speak for the Windows code base, obviously, but I do know that with the Linux and OpenSource ports to Qualcomm's Centriq line of Server chips it wasn't enough for us to just re-compile the code. Most compilers and source code have decades of optimization geared towards x86 architectures, putting entirely new platforms at a bit of a disadvantage.
    I meant specifically in comparing x86 version of App X running on WoA with some emulation, vs a native ARM compile of App X. Even a poorly compiled native version (GCC or such) would run circles around partial emulation.

    That's the biggest single factor in comparing ARM vs. x86 on Windows 10, getting a native version. It's pretty easy to do for UWP apps, but more challenging for Win32 programs. I suspect most open source projects will start adding ARM builds if WoA takes off a bit more with SD 850 hardware.
    Reply