Bytecode
Subject: Computer and Technology noun /ˈbaɪt.koʊd/ 1 definition
- compiler-design
- virtual-machines
- programming-languages
- runtime-environments
-
Definition (noun)
Entry #90 · UpdatedBytecode is a compact, platform-independent intermediate instruction set compiled from source code for execution by a virtual machine or managed runtime rather than directly by a physical processor. So named because its operation codes (opcodes) were historically one byte wide and paired with short operands, bytecode encodes low-level operations such as arithmetic, branching, memory access, and method dispatch in a dense binary stream. Unlike human-readable source code, it has already undergone parsing and front-end semantic checks to avoid repeated text processing at startup; unlike native machine code, it abstracts away target CPU registers and operating-system calling conventions, relying on an interpreter, Just-In-Time (JIT) compiler, or Ahead-Of-Time (AOT) compiler to run on host hardware.
Common Forms
Stack-Based Bytecode: Operates on an implicit evaluation stack where instructions pop operands and push results—used by the Java Virtual Machine (JVM), .NET Common Intermediate Language (CIL), CPython, and WebAssembly—producing compact instruction streams without explicit register addresses.
Register-Based Bytecode: Encodes virtual register indices directly inside each instruction to hold operands and intermediate values—used by Android's Dalvik Executable (DEX) format, Lua, and Erlang's BEAM—reducing total instruction count and stack-shuffling overhead at the cost of slightly wider instructions.
Typed (Verifiable) Bytecode: Embeds explicit type metadata and stack-frame descriptors so a runtime verifier can statically prove type safety, stack balance, and valid control-flow targets before execution, enforcing a security sandbox in JVM
.classfiles, .NET assemblies, and WebAssembly modules.Untyped (Dynamic) Bytecode: Emits generic, dynamically dispatched opcodes without static operand types—common in dynamic language runtimes such as CPython (
.pyc), Ruby (YARV), and V8 Ignition—deferring type specialization to runtime inline caches and adaptive JIT compilation tiers.
Example
When a developer compiles
InvoiceCalculator.javawithjavac, the compiler emitsInvoiceCalculator.classcontaining JVM bytecode instructions such asiload_1,iadd, andireturnrather than x86 or ARM machine instructions. That identical.classfile runs unmodified on both an Apple Silicon laptop and a Linux x86-64 production server because each host's JVM verifies, interprets, and JIT-compiles the bytecode into local native code at runtime.