The ROCm Revisited Series#
The ROCm Revisited series aims to revisit key concepts of the AMD ROCm software platform, tools, and optimizations, tailored for beginner and intermediate developers. This series shares our journey through the evolution of ROCm, highlighting the milestones, innovative technologies, and challenges we’ve overcome to establish leadership in the supercomputing space. Each post explores different aspects of ROCm’s development, focusing on how it has transformed industries, particularly in AI, machine learning, and high-performance computing (HPC). Through these blog posts, we’ll also discuss our commitment to open-source development and the future potential of distributed and energy-efficient computing. Below are the three blogs included in the series:
Powering the World’s Fastest Supercomputers#
Discover how AMD’s ROCm powers the world’s fastest supercomputers (El Capitan at Lawrence Livermore National Laboratory and Frontier at Oak Ridge National Laboratory). El Capitan leads the globe with over 1.7 exaflops of performance, while Frontier made history as the first system to break the exascale barrier. Both supercomputers harness the combined power of AMD Instinct GPUs, EPYC CPUs, and the ROCm software stack to deliver unmatched speed and efficiency.
ROCm’s open-source suite of tools, libraries, and frameworks is not just powering machines, it is enabling groundbreaking scientific discoveries, from simulating the universe to accelerating COVID-19 research. This technology is driving innovation, collaboration, and sustainability across the global HPC community, and is now a cornerstone for many of the world’s top supercomputers.
Dive deeper into how ROCm is shaping the future of engineering, science, and AI/ML. Explore the full story and see how ROCm enables us to push the boundaries of what’s possible.
Read the full blog: AMD ROCm: Powering the World’s Fastest Supercomputers
The High-Performance Computing Ecosystem#
Explore the journey of AMD’s ROCm ecosystem as it evolves from its early days into a powerful platform driving HPC, AI, and data science. Discover how ROCm’s compiler and runtime stack have advanced, highlighting the pivotal shift from HCC to HIP for enhanced compatibility and performance, and how its expanding libraries are accelerating machine learning, communication, and mathematical workloads. Dive into the latest developer tools for profiling, debugging, and system management, now optimized for multi-GPU and cluster environments. Whether you’re deploying on bare metal, in virtualized setups, or using containers, ROCm offers unmatched flexibility. The ecosystem now empowers a diverse range of applications, from leading AI frameworks like JAX and PyTorch to cutting-edge domains such as data science and simulation. Ready to see how ROCm can transform your projects? Read on to unlock the full potential of this dynamic platform.
Read the full blog: The High-Performance Computing Ecosystem
Getting Started with HIP Programming#
Are you ready to unlock the power of GPU programming? This article is your gateway to mastering HIP (Heterogeneous-compute Interface for Portability), AMD’s open-source API runtime designed for performance and portability. Whether you’re new to parallel computing or returning, this guide offers step-by-step guidance on core HIP concepts. Engaging examples and practical tips will help you quickly write, build, and run your first HIP programs; no prior experience required. Discover how HIP makes it easy to port CUDA applications across diverse hardware platforms, expanding your reach and impact. Dive in now and start your journey toward high-performance, cross-platform GPU development with HIP.
Read the full blog: Getting Started with HIP Programming
Summary#
In this brief “ROCm Revisited Series Navigation” blog we introduced you to our three-part series on the evolution, capabilities, and accessibility of AMD’s ROCm software platform. Each blog in the series dives into a distinct area: The history and future of open-source GPU supercomputing; the ROCm ecosystem, including its growth and optimizations for HPC; and an accessible guide to getting started with HIP programming. We hope that this post will serve as a useful navigation tool, summarizing the key takeaways from each of the series’ posts, enable easy access to the detailed blogs, and allowing you to explore and better understand the full potential of the ROCm platform.