# ROCm/hip

HIP: C++ Heterogeneous-Compute Interface for Portability

Repository: https://github.com/ROCm/hip
Canonical: https://ross.abutalabs.com/products/hip
Homepage: https://rocmdocs.amd.com/projects/HIP/
Language: C++
License: MIT
License Family: permissive
Topics: hip, hip-runtime, hip-portability, hip-kernel-language, hipify, cuda
Last push: 2026-08-25T08:22:59+00:00

## Health v2 (maintenance only)
Score: 95/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 86, longevity 100
- inputs: {"age_days": 3891, "days_push": 8, "days_rel": 96, "gap_med": 22.5, "n_releases_24m": 23}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4392, forks 592 (observed 2026-08-28T04:08:48.125397+00:00)

## What it is
HIP is a C++ runtime API and kernel programming language for AMD GPUs that mirrors the NVIDIA CUDA programming interface. It enables developers to write portable GPU code and automatically port existing CUDA applications to AMD hardware, primarily through the HIPIFY conversion tools.

## Use cases
- port existing CUDA code to run on AMD GPUs
- write GPU kernels in C++ for AMD hardware
- convert CUDA source files to HIP automatically
- build cross-vendor GPU-accelerated applications
- develop GPU compute workloads on the ROCm platform
- migrate CUDA-based machine learning code to AMD GPUs

## When to choose
- You need to migrate an existing NVIDIA CUDA codebase to run on AMD GPUs with minimal code changes
- You want a CUDA-like C++ API and kernel language for portable GPU programming
- You are developing GPU-accelerated applications targeting ROCm-supported AMD hardware on Linux or Windows

## When to avoid
- You exclusively target NVIDIA GPUs and will never need AMD support - using CUDA directly is simpler
- You rely on CUDA features or APIs that HIP does not support - check the CUDA-to-HIP API comparison first
- You need the full HIP runtime API on Windows, where some functions are not yet supported

## Facets
- artifact type: library
- maturity: active
- function: gpu-computing, sdk, programming-language
- domain: gpu-computing, machine-learning, deep-learning, developer-tools
- platform: windows, cpp
- tags: hip, rocm, amd-gpu, cuda-portability, hipify, gpgpu, kernel-language, heterogeneous-computing, cuda-migration, linux, gpu

## Member repositories
- ROCm/hip (main) score 95

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:48.125397+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:21:11.562242+00:00, confidence not recorded.
  - readme: https://github.com/ROCm/hip (fetched 2026-08-28T04:08:48.125397+00:00, sha ea0d8db5ea84)
  - homepage: https://rocmdocs.amd.com/projects/HIP/ (fetched 2026-08-29T09:09:13.147605+00:00, sha 04fc1b77707d)
  - site_page: https://rocm.docs.amd.com/projects/HIP/en/latest/faq.html (fetched 2026-08-29T09:09:13.156703+00:00, sha 9a6971429dc8)
  - site_page: https://rocm.docs.amd.com/projects/HIP/en/latest/install/install.html (fetched 2026-08-29T09:09:13.158880+00:00, sha ca22987ddbf6)
  - site_page: https://rocm.docs.amd.com/projects/HIP/en/latest/install/build.html (fetched 2026-08-29T09:09:13.160553+00:00, sha 4b50cafb822c)
  - site_page: https://rocm.docs.amd.com/projects/HIP/en/latest/reference/hardware_features.html (fetched 2026-08-29T09:09:13.162300+00:00, sha e52c053e8453)
  - site_page: https://rocm.docs.amd.com/en/latest/contribute/contributing.html (fetched 2026-08-29T09:09:13.163910+00:00, sha b6f13d8c7cc2)
  - site_page: https://rocm.docs.amd.com/en/latest/about/license.html (fetched 2026-08-29T09:09:13.165555+00:00, sha 635f4a453d95)
- Data as of 2026-08-30T08:39:29.467469+00:00.