Just-in-time compilation

Just-in-time compilation translates program code into machine code while a program runs. It can use runtime observations to optimize frequently executed code.

Interpreter: An interpreter executes program instructions without first translating the whole program into machine code. JITs often begin by interpreting code, then compile sections that run frequently.
Machine code: Processor instructions encoded in a form that a particular computer architecture can execute directly. A JIT ultimately emits machine code for the running program.
Ahead-of-time compilation: Compilation that translates a program before execution, typically producing a deployable native executable or object code. AOT moves compilation cost out of runtime, unlike JIT compilation.
Java HotSpot: The Java virtual machine implementation that uses tiered compilation and runtime profiling to optimize Java code. HotSpot is a prominent example of a production runtime JIT.
Warm-up period (computing): The initial phase in which a program's runtime gathers data and performs work before reaching steadier performance. Compilation and profiling can make JIT-compiled programs slower at first than after warm-up.
Hot spot (compiler optimization): A frequently executed program region selected for special optimization or compilation. Runtime counters help a JIT identify code worth compiling.
Compiler: A program that translates source code or an intermediate representation into another form of code. JIT compilation is compilation scheduled during program execution.
Static compilation: Compilation that resolves program structure and operations before execution rather than relying on runtime adaptation. Static compilation generally has less runtime information for optimization than a JIT.
V8: Google's JavaScript and WebAssembly engine, used in Chrome and other software, with optimizing compilation tiers. V8 demonstrates how browser engines compile dynamically typed JavaScript during execution.
Compilation overhead: The time and computing resources spent translating and optimizing code rather than executing its intended task. A JIT spends runtime CPU and memory to improve later execution speed.

Linked from 12 pages