User Guides

NOTE: If you are a user who is only interested in using an LLVM-based compiler, you should look into Clang instead. The documentation here is intended for users who have a need to work with the intermediate LLVM representation.

Clang

• How To Build On ARM

  Notes on building and testing LLVM/Clang on ARM.

• How To Build Clang and LLVM with Profile-Guided Optimizations

  Notes on building LLVM/Clang with PGO.

• How to cross-compile Clang/LLVM using Clang/LLVM

  Notes on cross-building and testing LLVM/Clang.

• How to build the C, C++, ObjC, and ObjC++ front end

  Instructions for building the clang front-end from source.

• LLVM Code Coverage Mapping Format

  This describes the format and encoding used for LLVM’s code coverage mapping.

• Control Flow Verification Tool Design Document

  A description of the verification tool for Control Flow Integrity.

LLVM Builds and Distributions

• Building a Distribution of LLVM

  A best-practices guide for using LLVM’s CMake build system to package and distribute LLVM-based tools.

• Building LLVM with CMake

  An addendum to the main Getting Started guide for those using the CMake build system.

• A guide to Dockerfiles for building LLVM

  A reference for using Dockerfiles provided with LLVM.

• Support Library

  This document describes the LLVM Support Library (lib/Support) and how to keep LLVM source code portable.

• Advanced Build Configurations

  This document describes more advanced build configurations.

Optimizations

• Writing an LLVM Pass

  Information on how to write LLVM transformations under the new pass manager.

• Writing an LLVM Pass (legacy PM version)

  Information on how to write LLVM transformations and analyses under the legacy pass manager.

• LLVM’s Analysis and Transform Passes

  A list of optimizations and analyses implemented in LLVM.

• Stack Safety Analysis

  This document describes the design of the stack safety analysis of local variables.

• MergeFunctions pass, how it works

  Describes functions merging optimization.

• LLVM Alias Analysis Infrastructure

  Information on how to write a new alias analysis implementation or how to use existing analyses.

• MemorySSA

  Information about the MemorySSA utility in LLVM, as well as how to use it.

• MemProf: Memory Profiling for LLVM

  User guide and internals of MemProf, profile guided optimizations for memory.

• LLVM Loop Terminology (and Canonical Forms)

  A document describing Loops and associated terms as used in LLVM.

• LLVM Cycle Terminology

  A document describing cycles as a generalization of loops.

• Auto-Vectorization in LLVM

  This document describes the current status of vectorization in LLVM.

• LLVM Link Time Optimization: Design and Implementation

  This document describes the interface between LLVM intermodular optimizer and the linker and its design

• DTLTO

  This document describes the DTLTO implementation, which allows for distributing ThinLTO backend compilations without requiring support from the build system.

• The LLVM gold plugin

  How to build your programs with link-time optimization on Linux.

• Remarks

  A reference on the implementation of remarks in LLVM.

• Source Level Debugging with LLVM

  This document describes the design and philosophy behind the LLVM source-level debugger.

• How to Update Debug Info

  This document specifies how to correctly update debug info in various kinds of code transformations.

• Instruction referencing for debug info

  This document explains how LLVM uses value tracking, or instruction referencing, to determine variable locations for debug info in the final stages of compilation.

• Debug info migration: From intrinsics to records

  This is a migration guide describing how to move from debug info using intrinsics such as dbg.value to using the non-instruction DbgRecord object.

• Key Instructions debug info in LLVM and Clang

  This document explains how the debug info feature Key Instructions is implemented in LLVM.

• Instrumentation Profile Format

  This document explains two binary formats of instrumentation-based profiles.

• InstCombine contributor guide

  This document specifies guidelines for contributions for InstCombine and related passes.

• The LLVM Instrumentor Pass

  A comprehensive guide to the highly configurable Instrumentor pass for custom program instrumentation, including the interactive configuration wizard.

Code Generation

• Writing an LLVM Backend

  Information on how to write LLVM backends for machine targets.

• The LLVM Target-Independent Code Generator

  The design and implementation of the LLVM code generator. Useful if you are working on retargetting LLVM to a new architecture, designing a new codegen pass, or enhancing existing components.

• TableGen

  Describes the TableGen tool, which is used heavily by the LLVM code generator.

GlobalISel

• MIRPatterns

  Describes the design of MIR Patterns and how to use them.

JIT

• MCJIT Design and Implementation

  Describes the inner workings of MCJIT execution engine.

• ORC Design and Implementation

  Describes the design and implementation of the ORC APIs, including some usage examples, and a guide for users transitioning from ORCv1 to ORCv2.

• JITLink and ORC’s ObjectLinkingLayer

  Describes the design and APIs for the JITLink library, ORC’s new JIT linker.

• Debugging JIT-ed Code

  How to debug JITed code with GDB.

Additional Topics

• CommandLine 2.0 Library Manual

  Provides information on using the command line parsing library.

• Extending LLVM: Adding instructions, intrinsics, types, etc.

  Look here to see how to add instructions and intrinsics to LLVM.

• How To Add A Constrained Floating-Point Intrinsic

  Gives the steps necessary when adding a new constrained math intrinsic to LLVM.

• How to build Windows Itanium applications.

  Notes on assembling a Windows Itanium environment.

• How to Cross Compile Compiler-rt Builtins For Arm

  Notes on cross-building and testing the compiler-rt builtins for Arm.

• Using ARM NEON instructions in big-endian mode

  LLVM’s support for generating NEON instructions on big endian ARM targets is somewhat nonintuitive. This document explains the implementation and rationale.

• Support for AArch64 Scalable Matrix Extension in LLVM

  LLVM’s support for AArch64 SME ACLE and ABI.

• Compiling CUDA with clang

  LLVM support for CUDA.

• User Guide for NVPTX Back-end

  This document describes using the NVPTX backend to compile GPU kernels.

• User Guide for AMDGPU Backend

  This document describes using the AMDGPU backend to compile GPU kernels.

• AMDGPU Asynchronous Operations

  Builtins for invoking asynchronous data transfer operations in AMD GPUs.

• AMDGPU Memory Model

  This document describes AMDGPU Memory Model which overrides the LLVM memory model when a program is compiled for the AMDGPU target.

• DWARF Extensions For Heterogeneous Debugging

  This document describes DWARF extensions to support heterogeneous debugging for targets such as the AMDGPU backend.

• Allow Location Descriptions on the DWARF Expression Stack

  This document describes a DWARF extension to allow location descriptions on the DWARF expression stack. It is part of DWARF Extensions For Heterogeneous Debugging.

• AMDGPU Execution Synchronization

  This document describes how execution of threads can be synchronized on AMD GPUs.

• User Guide for SPIR-V Target

  This document describes using the SPIR-V target to compile GPU kernels.

• User Guide for the DirectX Target

  This document describes using the DirectX target to compile GPU code for the DirectX runtime.

• User Guide for RISC-V Target

  This document describes using the RISC-V target.

• RISC-V Vector Extension

  This document describes how the RISC-V Vector extension can be expressed in LLVM IR and how code is generated for it in the backend.

• Scheduling Information for RISC-V VCIX Instructions

  This document shows how the scheduling information for RISC-V’s XSfvcp extension – SiFive Vector Coprocessor Interface (VCIX) – works and how to customize them.

• Sandbox IR

  This document describes the design and usage of Sandbox IR, a transactional layer over LLVM IR.

• Telemetry framework in LLVM

  This document describes the Telemetry framework in LLVM.

• LFI

  This document describes the Lightweight Fault Isolation (LFI) target in LLVM.