Samsung released its first Windows laptop with an ARM processor last year, but like every other Windows on ARM device released to date it was powered by a Qualcomm Snapdragon processor.
The company’s next Windows on ARM PC could be something very different though – according to a report from ZDNet Korea, Samsung is planning to launch a Windows 10 PC with a Samsung Exynos processor featuring AMD graphics.
Samsung hasn’t made any official announcements yet, but ZDNet says it’s reporting is based on information from “major component suppliers and partners.”
Samsung announced in 2019 that it was licensing graphics technology from AMD for use in future products, possibly including smartphones. And that fueled speculation that Exynos chips with AMD Radeon graphics were on the way. But aside from the occasional rumor, we haven’t heard much about those chips yet.
Now ZDNet says the processor, which might be known as the Exynos 2200, should be ready in third quarter of 2021, which is when Samsung plans to debut the processor in its new Windows on ARM computer, which will most likely be a thin and light laptop similar to last year’s Samsung Galaxy Book S.
It’s unclear how Samsung’s new chip will stack up against Qualcomm’s Snapdragon 8cx Gen 2 processor, which is currently the most powerful chip used in Windows 10 on ARM computers. But hopefully a bit of competition in this space will drive Qualcomm and other chip makers to step up their game in terms of performance.
While Windows on ARM PCs tend to offer long battery life and support for cellular networks, they also tend to be rather sluggish compared to models with Intel or AMD chips when running some applications that aren’t designed to run natively on ARM architecture. That’s because there’s a fair amount of extra work that the chips need to do in order to translate x86 apps to ARM on the fly.
Apple gets around this with its new ARM-based chips for Macs by effectively translating legacy apps before they’re run, allowing many older programs to run faster on Macs with Apple Silicon than they do on Intel-powered Macs. But unless Microsoft finds a way to improve performance for running x86 Windows apps on ARM-based PCs, the next best move is probably for chip makers to offer ARM-based processors with better performance.
thanks Phil!
I’d love to compare an Exynos octa core powered laptop running ChromeOS, Ubuntu and Windows 10. My guess is bloated Windows 10 performance would come in last place.
I wonder how the performance will be. With Apple’s M1, the bar went higher. Qualcomm’s ARM SoCs for Windows couldn’t meet a lower bar even before the M1 came out.
Although, I’m only interested in Windows on ARM hoping it’ll make it into UMPCs and not just 14″+ notebooks.
If Samsung adds 16 GB of RAM on the SOC together with an AMD GPU, I expect the performance to be a lot closer to the M1. By Q3, we should also get Windows 21H2, hopefully the X64 emulation will have improved dramatically by then.
Actually Samsung should be more suitable at Tablet/Ultrabook processor than Qualcomm, and even moreso for Nvidia.
Why?
Snapdragon has beaten Exynos in the phone segment for many years, boasting performance that responds quicker, then clocks down, without wasting as much electricity and producing as much heat. Also Snapdragon SoC’s don’t require much optimisation from the ODM to get much out of the device.
However, at these larger device sizes, you get more surface area to dissipate heat, and the larger batter allows for higher voltage. So most of Samsung’s disadvantages kind of disappear. Yet, Samsung’s Large Cores Mongoose (M3-M6) can stretch it’s legs over the Medium Cores (Cortex A75-A78) of Qualcomm. Qualcomm finally has a Large Core competitor in the Cortex-X-A78 but their implementation is mediocre. This phenomenon also applies in the iGPU segment where Samsung’s use of newer and newer Mali GPU saw them basically catch up to Qualcomm’s Adreno performance lead. However, the Mali GPU does use more power and throttle sooner, two huge disadvantages on a small mobile phone device.
Nvidia, mostly absent from the market, does have a miniaturised/mobile version of their dGPU that’s more impressive but it’s even more thirstier and throttles than the Mali GPU. Also, Nvidia does have a Custom ARM design in their Large Cores (Denver-Carmel) which isn’t that great but it’s meaty enough to not bottleneck the system and keep it’s GPU/NPU coprocessors running at full bandwidth. I could see their implementation running very good on an “ARM-book”. Nonetheless, a Snapdragon-book requires least optimisation, then more so for an Exynos-book, and even more so for a Nvidia-book, and most optimsation required for a A14/mac-book.