Contents

• 1. Introduction
• 2. Getting Started
  • 2.1. System Requirements
  • 2.2. Installation
• 3. User Interface
  • 3.1. Error Handling
    • 3.1.1. Enumerations
    • 3.1.2. Functions
  • 3.2. General Information Query
    • 3.2.1. Functions
  • 3.3. Compilation
    • 3.3.1. Functions
    • 3.3.2. Typedefs
  • 3.4. Supported Compile Options
  • 3.5. Precompiled header (PCH) (CUDA 12.8+)
    • 3.5.1. Functions
  • 3.6. Host Helper
    • 3.6.1. Functions
  • 3.7. Bundled Headers Installation
    • 3.7.1. Macros
    • 3.7.2. Functions
• 4. Language
  • 4.1. Execution Space
  • 4.2. Separate Compilation
  • 4.3. Dynamic Parallelism
  • 4.4. Integer Size
  • 4.5. Include Syntax
  • 4.6. Predefined Macros
  • 4.7. Predefined Types
  • 4.8. Builtin Functions
  • 4.9. Default C++ Dialect
• 5. Basic Usage
• 6. Precompiled Headers (CUDA 12.8+)
  • 6.1. Overview
  • 6.2. Implementation overview
  • 6.3. Automatic PCH
  • 6.4. Explicit PCH Creation and Use
  • 6.5. Determining the header stop point
  • 6.6. PCH failure conditions
  • 6.7. PCH Heap controls
  • 6.8. Miscellaneous Controls
    • 6.8.1. Environment variables
    • 6.8.2. Pragmas
    • 6.8.3. Flags
    • 6.8.4. Template instantiation before creating PCH
• 7. Tile Programming (CUDA 13.3+)
• 8. Accessing Lowered Names
  • 8.1. Example
  • 8.2. Notes
• 9. Interfacing With Template Host Code
  • 9.1. Template Host Code Example
• 10. Versioning Scheme
  • 10.1. NVRTC Shared Library Versioning
  • 10.2. NVRTC-builtins Library
• 11. Caching (CUDA 12.9+)
• 12. Bundled Headers (CUDA 13.3+)
  • 12.1. Overview
  • 12.2. Installation and Usage
  • 12.3. Best Practices
  • 12.4. Known Limitations
• 13. Miscellaneous Notes
  • 13.1. Thread Safety
  • 13.2. Stack Size
  • 13.3. NVRTC Static Library
• 14. Example: SAXPY
  • 14.1. Code (saxpy.cpp)
  • 14.2. Saxpy Build Instructions
• 15. Example: Using Lowered Name
  • 15.1. Code (lowered-name.cpp)
  • 15.2. Lowered Name Build Instructions
• 16. Example: Using nvrtcGetTypeName
  • 16.1. Code (host-type-name.cpp)
  • 16.2. nvrtcGetTypeName Build Instructions
• 17. Example: Dynamic Parallelism
  • 17.1. Dynamic Parallelism Build Instructions
• 18. Example: Device LTO (link time optimization)
  • 18.1. Code (offline.cu)
  • 18.2. Code (online.cpp)
  • 18.3. Device LTO Build Instructions
• 19. Example: Automatic PCH (CUDA 12.8+)
  • 19.1. Code (auto_pch_common.h)
  • 19.2. Code (auto_pch.cpp)
  • 19.3. Automatic PCH Build Instructions
• 20. Example: Explicit PCH Create/Use (CUDA 12.8+)
  • 20.1. Code (common.h)
  • 20.2. Code (explicit_pch.cpp)
  • 20.3. Explicit PCH Build Instructions
• 21. Example: PCH Heap Resizing (CUDA 12.8+)
  • 21.1. Code (auto_pch_common.h)
  • 21.2. Code (pch_resize.cpp)
  • 21.3. PCH Heap Resizing Build Instructions
• 22. Example: Tile compilation (CUDA 13.3+)
  • 22.1. Code (nvrtc_tile_simt.cpp)
  • 22.2. Tile compilation Build Instructions
• 23. Example: Bundled Headers (CUDA 13.3+)
  • 23.1. Code (bundled_headers.cpp)
  • 23.2. Bundled Headers Build Instructions
  • 23.3. Notices
    • 23.3.1. Notice
    • 23.3.2. OpenCL
    • 23.3.3. Trademarks