LLVM

LLVM is a collection of modular and reusable compiler and toolchain technologies designed to support the development of frontends and backends for programming languages. At its core is the LLVM intermediate representation (LLVM IR), a language-agnostic, low-level, typed instruction sequence that enables optimization and analysis to be performed independently of a specific source language or target architecture.

The LLVM project provides a mature optimizer framework and code generation infrastructure that can translate LLVM IR into machine code for many processors. It also offers a family of tools and libraries for building compilers, JITs, and dynamic-language runtimes. Clang, the project’s most widely used frontend, compiles C, C++, and Objective-C; other frontends exist for languages such as Swift, Rust, and Julia. LLDB serves as the debugger, and libc++ provides a modern C++ library implementation used in many toolchains.

History: LLVM originated in the early 2000s as a research project at the University of Illinois at Urbana-Champaign led by Chris Lattner. It was released as open-source software and has grown into a widely adopted infrastructure for compiler development. The project employs a permissive open-source license and welcomes community contributions from industry and academia.

Applications: LLVM underpins production toolchains on major platforms (notably Apple’s macOS and iOS toolchains) and serves as a backend for many language implementations. It is used in native code compilation, analysis and optimization passes, and runtime systems, including just-in-time compilation for dynamic languages and GPU code generation through experimental backends.