Passive ARM Assembly Skills for Debugging, Optimization (and Hacking)

How does your C++ code actually run on Apple Silicon? Learn to read ARM assembly to debug, optimize, and understand the compiler's secrets.

#1about 2 minutes

Core concepts of ARM assembly language

Assembly language consists of instructions and operands like registers, immediate values, memory addresses, and labels.

#2about 2 minutes

Overview of ARM CPU architecture and memory

Modern ARM CPUs like the Apple M1 feature numerous 64-bit registers and a multi-level cache system to mitigate slow main memory access.

#3about 3 minutes

Implementing a Fibonacci sequence loop in assembly

A practical example demonstrates basic ARM instructions like MOV, ADD, SUB, CMP, and B.NE for arithmetic and control flow.

#4about 3 minutes

Comparing handwritten assembly to optimized compiler output

The compiler often generates more efficient code than a direct translation, such as converting an incrementing for-loop into a decrementing counter.

#5about 4 minutes

Managing local variables using the stack

The stack provides fast, temporary memory for local variables, managed by the stack pointer (SP) and accessed with store (STR) and load (LDR) instructions.

#6about 4 minutes

How function calls work in ARM assembly

Function calls use the Branch and Link (BL) instruction, which saves the return address in the link register (X30) and transfers control.

#7about 2 minutes

The ARM procedure call standard and ABI

The Application Binary Interface (ABI) defines strict rules for register usage, including which registers pass arguments, return values, and must be saved by the caller or callee.

#8about 4 minutes

Loading 64-bit pointers with 32-bit instructions

ARM uses a two-step process with ADRP and ADD instructions to load a full 64-bit address by first getting the memory page and then the offset within it.