HIPIFY is a set of tools that you can use to automatically translate CUDA source code into portable HIP C++.

This is a linker wrapper for LLD; like ld-wrapper, it wraps the linker to add any missing -rpath flags, and to detect any misuse of libraries outside of the store.

This package provides the run-time support library developed by the LLVM project for the OpenMP multi-theaded programming extension. This package notably provides libgomp.so, which is has a binary interface compatible with that of libgomp, the GNU Offloading and Multi Processing Library.

The "clang-runtime" library provides the implementations of run-time functions for C and C++ programs. It also provides header files that allow C and C++ source code to interface with the "sanitization" passes of the clang compiler. In LLVM this library is called "compiler-rt".

This repository includes the user-mode API interfaces used to interact with the ROCk driver.

The Heterogeneous Interface for Portability (HIP) framework is a C++ Runtime API and Kernel Language that allows developers to create portable applications for AMD and NVIDIA GPUs from single source code.

The Comgr library provides APIs for compiling and inspecting AMDGPU code objects.

ROCm-CMake is a collection of CMake modules for common build and development tasks within the ROCm project. It is therefore a build dependency for many of the libraries that comprise the ROCm platform. ROCm-CMake is not required for building libraries or programs that use ROCm; it is required for building some of the libraries that are a part of ROCm.

AMD Common Language Runtime contains source code for AMD's compute languages runtimes: HIP and OpenCL. This package is built for HIP only.

The HIP compiler driver (hipcc) is a compiler utility that will call clang and pass the appropriate include and library options for the target compiler and HIP infrastructure.

RCCL (pronounced "Rickle") is a stand-alone library of standard collective communication routines for GPUs, implementing all-reduce, all-gather, reduce, broadcast, reduce-scatter, gather, scatter, and all-to-all. There is also initial support for direct GPU-to-GPU send and receive operations. It has been optimized to achieve high bandwidth on platforms using PCIe, xGMI as well as networking using InfiniBand Verbs or TCP/IP sockets. RCCL supports an arbitrary number of GPUs installed in a single node or multiple nodes, and can be used in either single- or multi-process (e.g., MPI) applications.

hipCUB is a thin wrapper library on top of rocPRIM or CUB. You can use it to port a CUB project into HIP so you can use AMD hardware (and ROCm software).

hipSPARSE is a SPARSE marshalling library with multiple supported backends. It sits between your application and a 'worker' SPARSE library, where it marshals inputs to the backend library and marshals results to your application. hipSPARSE exports an interface that doesn't require the client to change, regardless of the chosen backend. Currently, hipSPARSE supports rocSPARSE and cuSPARSE backends.

rocALUTION is a sparse linear algebra library that can be used to explore fine-grained parallelism on top of the ROCm platform runtime and toolchains. Based on C++ and HIP, rocALUTION provides a portable, generic, and flexible design that allows seamless integration with other scientific software packages.

OpenFabrics Interfaces (OFI) is a framework focused on exporting fabric communication services to applications. OFI is best described as a collection of libraries and applications used to export fabric services. The key components of OFI are: application interfaces, provider libraries, kernel services, daemons, and test applications.

Libfabric is a core component of OFI. It is the library that defines and exports the user-space API of OFI, and is typically the only software that applications deal with directly. It works in conjunction with provider libraries, which are often integrated directly into libfabric.

rocSOLVER is a work-in-progress implementation of a subset of LAPACK functionality on the ROCm platform.

The rocPRIM is a header-only library providing HIP parallel primitives for developing performant GPU-accelerated code on the AMD ROCm platform.

hipBLAS is a Basic Linear Algebra Subprograms (BLAS) marshalling library with multiple supported backends. It sits between your application and a 'worker' BLAS library, where it marshals inputs to the backend library and marshals results to your application. hipBLAS exports an interface that doesn't require the client to change, regardless of the chosen backend. Currently, hipBLAS supports rocBLAS and cuBLAS backends.

rocThrust is a parallel algorithm library that has been ported to HIP and ROCm, which uses the rocPRIM library. The HIP-ported library works on HIP and ROCm software.

hipFFT exports an interface that doesn't require the client to change, regardless of the chosen backend. It sits between your application and the backend FFT library, where it marshals inputs to the backend and marshals results back to your application.

rocBLAS is the ROCm Basic Linear Algebra Subprograms (BLAS) library. rocBLAS is implemented in the HIP programming language and optimized for AMD GPUs.

The rocRAND project provides functions that generate pseudorandom and quasirandom numbers. The rocRAND library is implemented in the HIP programming language and optimized for AMD's latest discrete GPUs. It is designed to run on top of AMD's ROCm runtime, but it also works on CUDA-enabled GPUs.

ROCm tracer provides an API to provide functionality for registering the runtimes API callbacks and asynchronous activity records pool support.

Unified Communication X (UCX) provides an optimized communication layer for message passing (MPI), portable global address space (PGAS) languages and run-time support libraries, as well as RPC and data-centric applications.

UCX utilizes high-speed networks for inter-node communication, and shared memory mechanisms for efficient intra-node communication.