LuisaCompute documentation

LuisaCompute is a high-performance cross-platform computing framework for graphics and beyond.

Contents:

• Getting Started
  • Overview
  • Prerequisites
  • Building from Source
  • Your First Program
  • Core Concepts
  • Python Frontend
  • Next Steps
  • Troubleshooting
• Tutorials
  • Table of Contents
  • ShaderToy-Style Mandelbrot Set
  • Path Tracing Renderer
  • MPM Fluid Simulation
  • Conway’s Game of Life
  • Wave Equation Simulation
  • Image Processing Pipeline
  • N-Body Gravitational Simulation
  • Fire Particle System
  • Reaction-Diffusion Simulation
  • Voxel Ray Tracer
  • Black Hole Renderer
  • Summary
• Project Architecture
  • Overview
  • Frontend: The Embedded DSL
  • Middle-end: Runtime and Resources
  • Backend: Code Generation and Execution
  • IR (Intermediate Representation)
  • Shader Compilation Pipeline
  • Memory Management
  • Threading Model
  • Extension System
  • Build System Integration
  • Performance Considerations
  • Debugging and Profiling
  • Future Directions
• Domain Specific Language (DSL)
  • Overview
  • Basic Types
  • Variables and References
  • Control Flow
  • Multi-Stage Programming
  • Callables
  • Kernels
  • Built-in Functions
  • Automatic Differentiation
  • Shared Memory
  • Atomic Operations
  • Structure-of-Arrays (SOA)
  • Ray Queries
  • Runtime Polymorphism
  • Binding Groups
  • Local Arrays
  • Custom Structures
  • Best Practices
• Resources and Runtime
  • Overview
  • Context
  • Device
  • Buffers
  • Images
  • Volumes
  • Bindless Arrays
  • Acceleration Structures (Ray Tracing)
  • Streams
  • Events
  • Memory Model and RAII
  • Command Scheduling
  • Shaders
  • Swapchain
  • Sampler
  • Depth Buffer
  • Raster Pipeline
  • Best Practices
• API Reference
  • Core Types
  • Runtime
  • Resources
  • Pixel Formats and Sampling
  • Shader Compilation
  • Ray Tracing
  • DSL Types
  • DSL Built-in Functions
  • Automatic Differentiation
  • Custom Structures
  • Sugar Syntax
  • Logging
  • Utility Classes