AArch64 and Arm64 Assembly on Raspberry Pi: Learning Path¶

This learning path takes you from a first AArch64 program to functions, debugging, C++ integration, GPIO, and NEON SIMD. It is designed for 64-bit Raspberry Pi OS on Arm64-capable boards.

Quick answer

AArch64 is the 64-bit execution state and instruction set; Arm64 is the commonly used platform name in Linux tools and package repositories. On Raspberry Pi OS, confirm uname -m reports aarch64, then start with registers and the calling convention before optimization.

Check your environment¶

uname -m
gcc --version
as --version
gdb --version

If uname -m reports armv7l, you are running a 32-bit userspace and these AArch64 examples will not assemble as written. Install the basic native toolchain with sudo apt install build-essential gdb.

Term	Meaning in this series
AArch64	Arm's 64-bit instruction set and execution state
Arm64	Linux/platform name normally corresponding to AArch64
AAPCS64	Procedure-call rules for arguments, results, registers, and stack
NEON/Advanced SIMD	Vector instructions using the `V0`–`V31` registers

Assembly tutorials¶

1. Arm64 Assembly on Raspberry Pi: AArch64 Tutorial¶

Verify a 64-bit Raspberry Pi OS environment, use GNU assembler (as) and linker (ld), run a Linux AArch64 hello-world program, and debug it with GDB.

2. ARM64 Registers and AArch64 Instructions¶

Learn the core register model: general-purpose registers (X0–X30, SP, ZR), AAPCS64 rules, floating-point registers, arithmetic, logical operations, and memory access.

3. ADRP, ADD, ADR, and LDR Literal¶

Learn how AArch64 constructs PC-relative addresses, loads nearby data, uses relocation pairs, and accesses the Global Offset Table.

4. Control Flow and Branching¶

Implement logic structures in assembly. Learn how processor flags, conditional branch instructions (B.cond, CBZ), and loops control execution flow.

5. Functions and Stack Management¶

Understand the AAPCS64 calling convention. Learn how parameters are passed via registers, manage the stack frame (SP), handle link register (LR) storage, and build nested functions.

6. Interfacing with C++ and GPIO¶

Write mixed-language applications. Call assembly routines directly from C++ source files, pass data across boundaries, and control physical Raspberry Pi GPIO pins using low-level memory maps.

7. NEON SIMD Vector Programming¶

Harness parallel computing. Learn how to use NEON vector registers and instructions to process multiple data elements simultaneously, accelerating image processing pipelines and math solvers.

8. Inline Assembly in C++¶

Embed raw assembly blocks directly inside your C++ code. Master GCC extended asm syntax, input/output constraints, clobber lists, and low-level instruction mapping.

Choose a route by goal¶

Goal	Read first	Continue with
Understand disassembly	Introduction and registers	Address loading and control flow
Call assembly from C/C++	Registers and functions	C++ integration and inline assembly
Debug crashes	Control flow and functions	GDB examples in each article
Optimize numeric code	Functions and ABI	NEON SIMD, then measurement
Access hardware	C++ integration	GPIO and kernel-supported interfaces

A repeatable practice loop¶

For every example, keep the source, exact build command, disassembly, and observed output together. Compile with debug symbols, inspect the generated instructions, single-step the code, then benchmark only after verifying correctness:

gcc -g -O0 example.s -o example
objdump -d example | less
gdb ./example

Optimization claims should include the compiler flags, input size, number of runs, CPU temperature, and throttling state. Use the reproducible Raspberry Pi benchmark guide when comparing assembly with compiler-generated code.