Composable Kernel is a library which provides a tile-based programming model for writing performance-portable kernels, in particular for GPUs using the HIP programming language.

This package contains a wrapper library for generating random numbers on GPUs, in particular via rocRAND for AMD GPUs.

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

This package contains a library for generating and analyzing AMD GPU assembly kernels.

This package contains a wrapper library for computing fast Fourier transforms on GPUs, in particular via rocFFT for AMD GPUs.

MIOpen is a library which exposes a C API to GPU machine learning primitives. This package contains the HIP version based on ROCm.

This package contains a library for generating and analyzing AMD GPU assembly kernels.

origami provides a method for selecting GEMM configurations deterministically based on compute and memory latencies.

This package contains a wrapper library for linear algebra algorithms on GPUs, in particular via rocRAND, rocBLAS, rocSPARSE and rocSOLVER for AMD GPUs.

rocThrust is a parallel algorithm library for ROCm, based on rocPRIM. It is a port of the CUDA Thrust library.

This package contains a HIP library for computing fast Fourier transforms.

This package exposes a sparse BLAS interface for ROCm. It is implemented in the HIP programming language.

This package contains a wrapper library with HIP parallel primitives, in particular via rocPRIM for AMD GPUs.

This package provides a header-only C++ library for a half-precision floating point type (IEEE 754 conformant) together with common operations on this type.

This package contains implementations for pseudorandom and quasirandom number generation in HIP.

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

User-mode API interfaces and libraries necessary for host applications to launch compute kernels to available HSA ROCm kernel agents.

HIP language runtime, allowing execution of HIP kernels on AMD harware, with library support for in-process/in-memory compilation (hipclr) of HIP kernel code, and its execution on AMD hardware.

AQLprofile is an open source library that enables advanced GPU profiling and tracing on AMD platforms.

The AMD SMI allows managing and monitoring AMD devices, particularly in high-performance computing environments. It provides a user-space interface that allows applications to control GPU operations, monitor performance, and retrieve information about the system's drivers and GPUs. It also provides a command-line tool, amd-smi, which can be used to do the same.

RCCL (ROCm Communication Collectives Library) is a library for collective communication on GPUs, such as reduce, gather or scatter operations.

RocBandwidthTest is designed to capture the performance characteristics of buffer copying and kernel read/write operations. The help screen of the benchmark shows various options one can use in initiating cop/read/writer operations. In addition one can also query the topology of the system in terms of memory pools and their agents.

The ROCTracer library provides APIs for tracing (also asynchronous) runtime calls in an application. It is intended for tracing ROCm API calls in GPU applications, such as kernel dispatches and memory moves.

The ROCm System Management Interface Library, or ROCm SMI library, is part of the Radeon Open Compute ROCm software stack. It is a C library for Linux that provides a user space interface for applications to monitor and control GPU applications.