Qualcomm’s new Snapdragon 850 processor is designed specifically for Windows on ARM devices like the Lenovo Yoga C630 and Samsung Galaxy Book2.
The chip maker says it offers up to 30 percent better performance over the Snapdragon 835 processor that powered earlier devices like the Asus NovaGo… which is a good thing, because the NovaGo is really, really slow.
So is a 30 percent performance boost enough? The first Samsung Galaxy Book2 reviews are in and the consensus seems to be… maybe? But probably not.
For example, Engadget found that the Galaxy Book2 managed to perform a batch image processing job in 2.5 minutes. The same task took about 3.5 minutes on the NovaGo. Progress, right? The only problem? The same job took about 10 seconds on recent devices with Intel processors.
That’s a tough pill to swallow when you’re looking at a 2-in-1 Windows tablet with a $1000 price tag.
Interestingly, PCWorld found that the Galaxy Book2 outperforms the Microsoft Surface Go with its Intel Pentium 4415Y processor in some web benchmarks… by just a little bit. But that’s a tablet that sells for $399 and up. And the Surface Go comes out way ahead when you run more general-purpose benchmarks like PCMark 8.
Laptop Magazine ran a number of additional benchmarks and for the most part they confirm that the Galaxy Book2 is a step up from Snapdragon 835-powered computers like the Asus NovaGo and HP Envy x2. But Samsung’s $1000 tablet doesn’t come close to matching Microsoft’s recently launched Surface Pro 6, which has an Intel Core i5 processor and a starting price of $899.
There are still some advantages to choosing a device with a Qualcomm processor rather than an Intel chip. The Galaxy Book2 is a fanless 12 inch tablet that weighs just 1.8 pounds and offers around 15+ hours of battery life in real-world tests. It also features built-in support for 4G LTE.
Samsung’s Galaxy Book2 features a 2160 x 1440 pixel AMOLED display, 4GB of RAM, 128GB of PCIe NVMe solid state storage, two USB 3.1 Type-C ports, a headset jack, a microSD card reader and a nano SIM card slot.
It features a built-in kickstand and comes with a Samsung S-Pen and detachable keyboard with a touchpad. Reviewers largely agree that the display looks great, the keyboard is pretty good, and overall built quality is stellar. The only real issue that keeps the Galaxy Book2 from competing with Microsoft’s Surface devices is the subpar performance when running resource-intensive apps and games.
The Book2 ships with Windows 10 Home in S Mode, which means you’re restricted to running apps downloaded from the Microsoft Store. But you can switch to the full version of Windows 10 Home for free — you just may notice that performance with non-Store apps isn’t quite as good since any app that’s not optimized for ARM architecture will require x86 emulation to run properly, and that adds a bit of overhead.
Even if you’re running the full version of Windows 10, you’ll have some restrictions on the types of x86 apps you can run: there’s no support for 64-bit x86 applications, so if you want to run software that doesn’t offer a 32-bit version you’re out of luck.
If you don’t need an always-connected, always-on thin and light computer with mobile broadband support, you could probably spend a lot less money and get a faster machine. But if you’re willing to put up with substandard performance in exchange for those features, I guess it’s nice to have the option. It would be even nicer if you didn’t have to choose between one and the other.
I imagine the big difference is that x86 chips benefit from years of optimisation and hardware-accelerated boost for windows, so an arm chip of equal processing power will always give much slower real world performance than an x86 chip.
The questions will be: Is the arm chip priced cheaply enough to compete? , and Can its battery greater life compensate for lack of processing speed?
It includes a pen and typecover, which would normally be $200-250 on other detachable so like the Surface Pro, iPad pro, and pixel slate.
So think of the pricing as $750-800, in line with a flagship phone, since these use a more powerful version of same processors, and similar premium build quality.
The x86 translation is a stop gap measure, the real comparison in power comes with native ARM64 apps.
Qualcomm is working on a ARM64 compilation of Chromium browser. I think Firefox was rumored to create an ARM64 version also. MS Store just got support for ARM64 in the latest 1809 windows update. And also got support for MSIX, which makes distribution of both ARM64 and x86-64 versions of both win32 and UWP apps within and outside of the store, quite seamless.
The only thing x86 benefits from is the fact that people create software for it. The notion that “an Arm of equal performance is slower” is BS. Remember, these Arm chips are emulating x86 instructions.
Im confident that if every piece of software out there was rewritten for Arm, most of them would be equal or better on Arm.
Since of them would be better on x86. There are some instruction types on x86 that are unique to Intel or AMD. Virtualization is one. Arm’s virtualization is in infancy, and isn’t comparable to Intel’s yet.
Why does anyone care about native LTE in a laptop? Every phone in the world has Hotspot
For the same reason someone would care about LTE in an iPad or Samsung Galaxy tab….
This is a tablet as much as a laptop.
So the takeaway is it is a great device…with the wrong architecture.
How do Qualcomm’s ARM SoCs compare with other vendors? Just wondering if it’s just Qualcomm that has slow SoCs or does ARM still have a ways before being able to compete with Intel in the desktop PC (at least “premium” notebook) space.
At least based on hearsay (ie. I have no data), Apple’s SoCs are “very fast” compared to other ARM based chips. Too bad they’ll never make it outside of Apple devices.
Wait for the A12X powered macbook that is coming end of next year. It will not disappoint.
Qualcomm is the new Intel.
Macs are rumored to transition to ARM in 2020 Ir 2021.
Assuming they start with the MacBooks which would be the most logical. The first model will have a derivative of the A13X not A12X which will be revealed on Tuesday.
Hi Brad – headline should read “…step in the “right” direction” or maybe “baby step in the right direction”
That was a typo… but I kind of like that it was unintentionally vague 🙂
Comments are closed.