x86 vs ARM: What’s The Difference?

x86 vs ARM: What’s The Difference?

Apple M1 is possibly the most familiar name among the general consumer crowd when it comes to CPUs in general, at least in recent times. Meanwhile, Windows PC users has always been using some variants of chips from either Intel or AMD, and even Apple once used them for many years before it gets dropped to make way for their in-house M-series chips.

While Apple claims huge performance numbers and impressive power efficiency from their own chips, there’s a crucial difference in between the chips you may be using right now, to what Apple is currently using for their Mac products: Instruction Set Architecture (ISA). Specifically, x86 and ARM, which is what most Windows PCs and Macbooks use, respectively. What does this mean?

x86: The Origin Of Modern PC

The x86 architecture, used in virtually every single Intel and AMD chip ever made in the past few decades, traces its history back to their “ancestor”: Intel 8086. This venerable little chip is launched all the way back in 1978, or 45 years ago. Its advantage in programming compatibility with previous architectures has made it popular amongst other more advanced ISAs at the time (including Zilog and Motorola), and the scale continues to grow larger and larger thanks to the growth in PC segment seen in the majority of 80s and 90s.

Intel 8086 – the very first x86 CPU.

That advantage still holds today – and there’s a reason for it. An instruction set architecture, like x86, is like a language for computers. It dictates how the most basic functions of a computer should work, for example, how a number should be calculated and how quickly it can be performed – like certain languages having its alphabetical system and phonetics dictating how a person should speak that language.

x86 is also what we call Complex Instruction Set Computer (CISC): in simple terms, it can do a lot of different maths at the same time, unlike ARM that is RISC – which we’ll explain later, so keep that in mind first.

ARM: Efficiency As The Purpose

ARM is what you’ll find inside virtually every single smartphone in your pocket right now. Why don’t they use x86 when it’s ubiquitous on PCs? The answer lies in the term you just saw: RISC, or Reduced Instruction Set Computer. To differentiate CISC and RISC, here’s an example.

Fun fact: ASUS was one of the few smartphone vendors that offered Intel Atom-powered devices, running on the x86 architecture. Intel’s stint in smartphone scene didn’t last long, however.

Imagine you need to buy things to set up your new house. Like, lots of it – tools, kitchen utensils, books, decorations, what-have-you. To do that, you’ll need to bring your whole family in a car, split up the team, and buy all the stuffs from all the different types of stores. You people puts all the stuffs into the car and drives back home. That’s essentially how x86, or CISC chips, work – a lot of things can be done in a single ride (or in this case, a compute cycle).

Meanwhile, RISC chips functions like this: imagine again, you just used up the toothpaste that brush your teeth twice a day with. You head to the store on your own with a small bike, buy the toothpaste, and get back home right away. You don’t need your whole family to tag along just to get one product. It’s quick – but you can only carry small amount of things each time you head to the store.

See the difference? CISC is good for complex compute workloads, but you may end up with a family member that just tags along and do not actually carry stuffs (like a part of chip that just sits there and not being utilized, which is not always efficient). RISC meanwhile is better for efficiency, where you don’t need to call up your whole family to complete a simple task.

Different “Language”

However, the important thing to note here is that x86 and ARM are not compatible to each other. Most apps you see today that runs on your Windows PC is designed based on the x86 architecture, and it simply will NOT run if the app is placed inside a laptop that is powered by an ARM chip (think Apple M1 or Qualcomm Snapdragon 8cx). Similarly, apps designed for ARM chips, can’t talk to PCs with x86 hardware in it – that is, every Intel and AMD-powered system out there.

Apple M1 and Apple M2.

App developers have to specifically design their apps to work on each architecture – think of it like writing an essay in two different languages, say – English and Chinese. Chinese speakers may not fully understand the essay that is written in English, and vice versa, so apps also needs to understand the “language” the PC is using. And just like Google Translate, a translation layer for two different ISAs exist: the more well-known one being Apple’s “Rosetta” which allows its M1/M2-powered devices to at least run the apps that is designed for the x86 architecture.

Apple Rosetta.

However, going back to the language analogy again, giving a Chinese native speaker a piece of English essay, even if the said person understands English, is still not optimal – since it’s not their native language. What that means is effectively a performance loss in the context of computers: which varies from slightly worse to significantly worse, depending on the hardware in question.

Should I consider an ARM-powered laptop next?

If you’re an Apple user, this question is largely irrelevant – you are likely already using one by now, unless you have a fairly old model. However, for Windows users, things can get a bit muddy. Windows has largely built its foundation on the x86 architecture, with the help of Intel (hence “Wintel” is a term coined in this industry). That’s nearly 40 years worth of hardware and code – and given the sheer amount of Windows PCs out there (and the size of the Windows ecosystem), it’s proven to be very difficult to transition to an architecture that effectively requires a “big reset”.

A Windows-on-ARM development kit for app developers to design apps on ARM-based PCs.

ARM Windows laptops do exist, but in very small numbers. Some ultra-lightweight laptops, often 2-in-1s, are powered by Qualcomm’s 8cx chips – think of it as the souped-up version of whatever’s being offered in the current flagship smartphones. However, at current state, it’s still quite limited for what it can do. You’ll basically have to use these laptops similarly to how you use phones, that is to say, apps that runs are pretty much coming from the Microsoft Store, and anything you downloaded from the Internet (which is often referred as Win32 apps) is very unlikely to work natively, or even at all.

Microsoft Surface Pro X – powered by Microsoft SQ2, a variation of Qualcomm’s chips.

There’s still a long way to go for ARM in Windows, so my current advice is, stick to x86, like Intel or AMD chips, if you work on many complex stuffs like 3D, video rendering and such. For users that who just needs something to do very basic tasks, like typing and documenting – the efficiency provided by ARM chips may outweigh the relative lack of versatility, at least for now. It remains to be seen if they can challenge x86’s current status quo in a few years, and whether app developers may finally catch up by then.

Info source: Red Hat | LowSpecGamer (YouTube) | TechQuickie (YouTube)