Estimated reading time: 18 minutes

Over the past years, we have seen a shift in processors from the previously dominant x86 architecture to the more energy-efficient (and often cheaper) ARM architecture. This trend is true for both consumer hardware, e.g., Apple’s M1 SoC and also server hardware, e.g. AWS’s Gravitron processor. Given that the ARM architecture has a more relaxed memory model, this might have some subtle impact on C++ programs…

With the introduction of multi-threaded execution in C++11, the standard also introduced a memory model. For many developers, this memory model is one of the least well-understood parts of the C++ standard and a common source of confusion. While an in-depth explanation of the C++ memory model is beyond this blog post’s scope, we will cover some basics and take a brief look at two of the most common hardware memory models and how the C++ memory model maps to these hardware models.

But first, let’s take a step back and consider why we need a memory model in the first place.