ROCm

ATMI is a runtime framework and programming model that enables performance portability across heterogeneous computing platforms.

A comprehensive tool suite for the performance-aware developer. It includes powerful debugging, comprehensive profiling, and frame-level and ISA-level analysis, enabling developers to harness the benefits of CPUs, GPUs and APUs.

HSA Acclerated Python via Numba by Continuum Analytics

HCC is a open source, optimizing C++ compiler for heterogeneous compute platforms.  It is based on clang and the LLVM compiler infrastructure.  It is optimized for high performance compute accelerators with Boltzmann software stack.  It also supports HSA capable APUs using the open source driver.

HIP (Heterogeneous-compute Interface for Portability) is a portable C++ runtime and kernel language for GPUs, and includes tools to "hipify" CUDA code into the portable C++ language.

HipSYCL - SYCL 1.2.1 Third Party Solution runs on HIP

An LLDB-based debugging environment for debugging both the host application and GPU kernels running on AMD HSA hardware.

OpenCL SDK for AMDGPUpro

“The Radeon Compute Profiler (RCP) is a performance analysis tool that gathers data from the API run-time and GPU for OpenCL™ and ROCm/HSA applications.”

Radeon GPU Analyzer is an offline compiler and performance analysis tool for DirectX® shaders, OpenGL® shaders, Vulkan® shaders and OpenCL™ kernels

ROCm GDB provides a gdb-based debugging environment for debugging host application and GPU kernels running on Radeon Open Compute platform.

Status Monitor System is an example of a multi GPU OpenCL™ program that exercises all or part of installed GPUs.

AMD OpenVX (beta preview) is a highly optimized open source implementation of the Khronos OpenVX™ computer vision specification. It allows for rapid prototyping as well as fast execution on a wide range of computer hardware.

ArrayFire is an open-source parallel-computing library that delivers high performance and offers an easy-to-use API. Its array-based functions make parallel programming simple. By providing multiple back ends, ArrayFire enables you to easily develop efficient, portable applications.

BLIS is a portable software framework for instantiating high-performance BLAS-like dense linear algebra libraries. It provides a super-set of the classic BLAS API, free from Fortran dependencies.

clBLAS is a software library containing the complete set of BLAS level 1, 2 and 3 routines. In addition to GPU devices, the library also supports running on CPU devices to facilitate debugging and multicore programming.

clFFT is a software library containing Fast Fourier Transform functions written in OpenCL. clFFT provides a set of FFT routines that are optimized for AMD graphics processors, but also are functional across CPU and other compute devices.

clRNG is an OpenCL™ based library for uniform random number generation. The library provides multiple streams, and substreams as needed, for parallel applications and simulations.

clSPARSE is an OpenCL™ library implementing Sparse linear algebra routines. It's currently in Beta, and provides a subset of sparse BLAS L2 and L3 operations, and a few iterative sparse solvers.

HCC implementation of Basic Linear Algebra Subroutines which is gpu accelerated of the standard BLAS library

HCC based Fast Fourier Transform library provides a simple interface for computing FFTs on GPU.

HCC enabled library for uniform random number generation

High-performance object-based library for DLA computations on host processors. It provides a super-set of the classic LAPACK API, free from Fortran dependencies.

MIOpen a deep learning performance solver library

Radeon Loom Stitching Library (beta preview) is a highly optimized library for 360 degree video stitching applications.

The AMD GCN ISA Architecture document describes the environment, organization, and program state of AMD GCN Generation 3 devices. It details the instruction set and the microcode formats native to this family of processors that are accessible to programmers and compilers.

Using FirePro DirectGMA, any PCI-E device can directly write to GPU memory. FirePro DirectGMA eliminates unnecessary system memory copies, dramatically lowers CPU overhead, and reduces latency, resulting in significant performance improvements in data transfer times.

The ROCm platform delivers on the vision of the Boltzmann Initiative, bringing new opportunities in GPU Computing. It is an open platform for development, discovery and education around GPU Computing.

CoMD is a reference implementation of typical classical molecular dynamics algorithms and workloads

HCC enabled Haccmk, Hardware Accelerated Cosmology Code (HACC) framework uses N-body techniques to simulate the formation of structure in collisionless fluids under the influence of gravity in an expanding universe

HPGMG implements full multigrid (FMG) algorithms using finite-volume and finite-element methods.

LULESH Computer simulations of a wide variety of science and engineering problems require modeling hydrodynamics, which describes the motion of materials relative to each other when subject to forces.

Radeon Loom Stitching Library (beta preview) is a highly optimized library for 360 degree video stitching applications.

SNAP (SN (Discrete Ordinates) Application Proxy) is a proxy application to model the performance of a modern discrete ordinates neutral particle transport application

XSBench is a mini-app representing a key computational kernel of the Monte Carlo neutronics application OpenMC