What is the difference between ARM architecture and the typical architecture in CPUs, and why is it so revolutionary?

960 viewsEngineeringOther

What is the difference between ARM architecture and the typical architecture in CPUs, and why is it so revolutionary?

In: Engineering

14 Answers

Anonymous 0 Comments

It’s not that revolutionary. It’s been used for various smaller chips for decades now.

However ARM’s business model is to be flexible with working with customers – if the customer wants them to design an entire CPU core, they’ll do that. If the customer just wants access to the ISA, they’ll do that. And varying levels of IP access in between. However, they do not actually sell any chips themselves.

This attracted a lot of mobile chip companies like Qualcomm and Samsung and Apple who wanted an accessible ISA as well as certain levels of IP access. Intel and AMD, the 2 main chipmakers at the time, had very difficult-to-access IP, and their off-the-shelf chips that they sold at the time were not very well suited for mobile applications due to high power consumption and wonky workarounds to maintain backwards compatibility.

Since ARM chips were primarily used for low power stuff like toys and sensors, they were better suited for mobile applications and didn’t have all the wonky workarounds.

Once the major mobile chipmakers were on board, software companies started programming for ARM chips, which makes it harder for non-ARM chips to break into the market since it means that all the software would have to be re-written.

Anonymous 0 Comments

First ARM has been around since the late 70s.  But at the time there were basically two main types of CPU architecture: complex instruction set computers (CISC), and reduced instruction set computers (RISC)

The idea behind CISC was to make a lot of different instructions to do common tasks that software might want.  RISC took the opposite approach: implement the bare minimum set of instructions, but use the simplicity to make the CPU run those instructions really fast.

Intel went down the CISC route, and ARM(Called Acorn at the time) chose RISC.  Over time, Intel kept adding more and more instructions and had to keep supporting all of them, whether or not they were widely used, or risk creating incompatibilities with older software.  ARM didn’t really have this problem, since they rarely added new instructions.  Over time, this bloat caught up with Intel/AMD.

One of the side effects of having a simple CPU instruction set, is that the circuits to power those instructions is so simple.  Simple circuits lead to fewer transistors, which ultimately leads to less power consumption.  While that may not be important on a desktop computer, it’s becoming more important as more of our devices are battery powered.  And on the data center side, cutting power consumption on a large scale leads to huge saving in both power and cooling.

And the x86 compatibility is becoming less of an issue these days as it’s super easy to port software to other architectures, or even run a whole x86 emulation layer in software.  And the power savings is dramatic as evidenced by the new ARM powered Macs. (M1/M2/etc)

Anonymous 0 Comments

ARM is more optimised for porn usage. The instruction set is heavily dedicated for acceleration of the amount of porn videos per second you can watch. All the porn sites leverage this to give you as many dicks per second.

Anonymous 0 Comments

Here is an ELI5 for you – I big truck can move a large load and takes quite a bit of fuel, a fleet of 4×4 can move the same load but only a 1/40 of the rate (to the same destination.

What makes a difference is where there are multiple different destinations or the amount varies so if you only need 2 pallets why send a big truck.

So power efficient, and faster when going to different locations

for some (most loads) there is overhead in chunking up into pallets sizes and reconstituting.

Now for the trick, the 4×4 can tow multiple trailers, making them very efficient indeed.

It’s the ability to have multiple simple cores good what they do, and the cores can be idle when not in use (standby or all running in a video call).

This contrast the x86 which is a huge truck can move (computer) anything

In this example a GPU is a moped and there are x10,000 and they move pizza about and so far better than arm and x86.