HIP documentation

HIP is a C++ runtime API and kernel language for AMD GPUs. It allows developers to create portable applications by providing a programming interface that is similar to NVIDIA CUDA, making it straightforward to port existing NVIDIA CUDA code to run on AMD hardware. For more information, see What is HIP?

Note

HIP API 7.0 introduces changes to make it align more closely with NVIDIA CUDA. These changes are incompatible with prior releases, and might require recompiling existing HIP applications for use with the ROCm 7.0 release. For more information, see HIP API 7.0 changes.

Installation instructions are available from:

• Installing HIP

• Building HIP from source

The HIP documentation is organized into the following categories:

Programming guide

• Introduction to the HIP programming model

• Understanding GPU performance

• Hardware implementation

• HIP compilers

• Performance guidelines

• Debugging with HIP

• Logging HIP activity

• Using HIP runtime API

• HIP C++ language extensions

• Kernel language C++ support

• Porting CUDA code to HIP

• Programming for HIP runtime compiler (RTC)

• AMD compute language runtimes (CLR)

Reference

• HIP runtime API

• HIP math API

• HIP complex math API

• HIP environment variables

• HIP error codes

• NVIDIA CUDA to HIP API Function Comparison

• List of deprecated APIs

• Low Precision Floating Point Types

• Hardware features

Tutorial

• HIP basic examples

• HIP examples

• SAXPY tutorial

• GPU programming patterns

• Reduction tutorial

• Cooperative groups tutorial

• HIP Graph API tutorial

Known issues are listed on the HIP GitHub repository.

To contribute features or functions to the HIP project, refer to Contributing to HIP. To contribute to the documentation, refer to Contributing to ROCm docs page.

You can find licensing information on the Licensing page.