Skip to content

What is Apple M1’s Unified Memory? How much do you need?

  • by
  • 5 min read

Apple unveiled its new iteration of Macbooks in November, much like it does every year, but this time around their laptops had something special on the inside. Unlike the previous generation of Macbooks that ran on Intel processors, the newer systems use a chip created by Apple, called M1, running the ARM architecture, which is a huge deal for the laptop industry as most of the systems use Intel or AMD chipsets running the x86 architecture.

Apart from using a different architecture altogether, Apple also introduced a unified memory architecture for its SoC, which provides faster performance on its ARM-based chips.

Although a lot is going on with the new Macbook, understanding its memory architecture is key as the memory on the system cannot be upgraded. So, if you are looking to get a deeper understanding of Apple’s unified memory architecture and how much of it you need, you have come to the right place.

Also read: How much RAM do you need in 2020?

Understanding memory usage

Before we get into what Apple does differently, it’s important to understand how generic systems utilise RAM. 

What is Apple M1's Unified Memory? How much do you need?
Source: Apple

If you have a system with a dedicated GPU, then both your GPU and CPU have different sets of memory and each processor uses its own RAM to get the data for computational purposes. Let’s look at an example to get a better understanding of how things happen under the hood.

Let’s say you are running a game on your machine, in such a scenario the CPU loads a set of instructions into the RAM to execute the game and then sends data to the RAM of the GPU. In such an architecture, there is a lot of data transfer between the memory of the GPU and the CPU, and while separate RAMs allow you to run bigger programs and games without overloading the CPU, these data transfers can reduce the performance of your system if the data is massive.

What is Apple M1's Unified Memory? How much do you need?
An overview of memory architecture in systems with a dedicated GPU

In case of an integrated GPU, both the CPU and GPU use the same RAM chip, but both of them use a different space in the memory to store their data. Therefore, even with an integrated GPU, the memory space for the CPU and the GPU are separate and data needs to copied between the two memory pools.

That said technologies like HUMA and shared virtual memory have tried to lower the processing overhead caused by this issue.

In addition to all this, the CPU is connected to the RAM using sockets in traditional systems, which is not the case for an Apple device using proprietary silicon.

Also read: How to fix the infamous OBS black screen?

Decoding unified memory

In the case of Apple silicon, the usage and placement of the memory is a little different when compared to x86 systems.

What is Apple M1's Unified Memory? How much do you need?
Source: Apple

First of all, Apple uses a SoC design for its processor and everything from the CPU to the neural engine are on the same silicon. The memory for the system is not a part of the same silicon, but it uses the same substrate. Due to this, data transfer can occur at faster speeds and the processor can use higher bandwidth to access data when compared to systems, which use a motherboard to connect to the RAM.

In addition to this, the unified memory architecture allows any part of the SoC to access the memory at a given time, and each component does not have an independent memory pool. This means that data does not have to be copied from the CPU’s memory to the GPU’s or visa versa, which improves the overall performance of the system by getting rid of the copying operations.

What is Apple M1's Unified Memory? How much do you need?
Apple’s new line-up, announced in November 2020, feature the new Apple M1 chipset

All the changes mentioned above are the reason for the high benchmark scores of the new Macbook. Its performance is so mind-boggling that the new Macbook Air scores over a million points on the Antutu benchmark.

Also read: What is the difference between GPU and FPGA?

How much memory do you need on a Mac running M1?

Currently, Apple offers systems with an 8GB RAM configuration or a 16GB configuration. Although this amount of memory might sound absurdly low for a power user, the 8GB machine is enough to take care of your daily needs without even ramping up the fan.

So if you find yourself browsing with multiple tabs on your system and running a word processor, the 8GB system will probably get the job done.

If you are a power user and find yourself compiling large projects on Xcode or Lightroom getting the 16GB variant is a better idea.

That said if you are someone who edits a lot of 8k video on your system or export huge 4k videos you should reconsider your decision of buying a Macbook with Apple silicon as you cannot use more than 16GB RAM on the system.

Also read: 720p vs 1080p vs 2k vs 4k vs 8k: Which display should you buy?

Nischay Khanna

Nischay Khanna

A tech enthusiast, driven by curiosity. A bibliophile who loves to travel. An Engineering graduate who loves to code and write about new technologies. Can't sustain without coffee. You can contact Nischay via email: