CHIP-8 Assembler Experiment

A small CHIP-8 assembler experiment written in Go. The project was built to practice instruction parsing, bit-level opcode encoding, and binary output rather than to provide a complete toolchain.

What it demonstrates

• Splitting assembly source into instructions and operands
• Dispatching mnemonics to encoding functions
• Building 16-bit CHIP-8 opcodes from register and immediate fields
• Writing each opcode as high and low bytes to a .ch8 file

Parsing and encoding

cmd/main.go reads the input file line by line, selects an encoder from the first token, and constructs the four-hex-digit opcode for that instruction. The encoded value is parsed as a 16-bit integer and emitted in big-endian byte order.

The current dispatcher includes CLS, RET, JP, CALL, SE, SNE, OR, AND, XOR, ADD, SUB, SHR, SUBN, SHL, RND, DRW, SKP, and SKNP. LD is recognized but not implemented.

This repository contains the assembler only. The separate chip-8 repository contains an independent Rust emulator experiment.

Build and run

Requirements: Go and Make.

make build
make run

The current entry point uses fixed project-relative paths:

• input: cmd/asm/first.asm
• output: cmd/out/output.ch8

It can also be run directly:

go run ./cmd

Limitations

• The assembler has no lexer, labels, symbols, macros, or command-line file arguments.
• Operand validation and error reporting are minimal.
• The instruction set is incomplete, including the unimplemented LD family.
• Parsing assumes a narrow whitespace and operand format.
• There is no automated conformance test suite.