PPPL aims to deploy a hybrid CPU-GPU-QPU architecture for accelerating the timeline for commercial fusion energy.
The Princeton Plasma Physics Laboratory has launched a computing platform titled Simulation, Technology, and Experiment Leveraging Learning-Accelerated Research enabled by AI, or STELLAR-AI. This initiative aims to address the computational delays that currently hinder fusion energy research by integrating artificial intelligence with high-performance computing.
Fusion research involves complex simulations of plasma behavior that can require several months to complete using existing infrastructure. “Fusion is a complex system of systems. We need AI and high performance computing to really optimize the design for economic construction and operation,” said Jonathan Menard, deputy director for research at PPPL in a press release.STELLAR-AI is designed to shorten these timelines by connecting computing resources directly to experimental devices. Hardware architecture enables integrationThis configuration allows for the analysis of data in real-time as experiments occur. A primary experimental connection for the platform is the National Spherical Torus Experiment-Upgrade at PPPL, which is scheduled to begin operations later this year.The hardware architecture of STELLAR-AI consists of a combination of central processing units , graphics processing units , and quantum processing units . CPUs are utilized for standard computing tasks, while GPUs provide the parallel processing capabilities necessary for training AI models. QPUs are included to solve specific complex calculations that traditional computers cannot process efficiently. This hardware configuration is intended to support the technical requirements of private fusion companies working to develop commercial power solutions.Under the framework of the Genesis Mission, STELLAR-AI contributes fusion-specific computer codes and scientific models to a broader national system. For commercialization of fusion power plantsThe project supports the DOE’s Fusion Science and Technology Roadmap, which outlines the development of an AI-Fusion Digital Convergence platform. The stated objectives of this roadmap include the commercialization of fusion power plants and the provision of energy for future computing and AI infrastructure.“One effort will create a digital twin of NSTX-U: a computer model that mirrors the physical machine so closely that scientists can test ideas virtually before running actual experiments,” added the press release.“Another project, called StellFoundry, uses AI to speed the design of stellarators, a type of fusion device with a twisted, pretzel-like shape that some scientists believe could offer advantages over other designs.”The design process for these machines involves analyzing large datasets to determine optimal configurations, a task the platform is intended to accelerate.A global fusion coalitionThe project is managed by PPPL and involves a network of national laboratories, academic institutions, and private companies. Partner laboratories include the UK Atomic Energy Authority , while academic participation includes Princeton University, the Massachusetts Institute of Technology, and the University of Wisconsin-Madison. Princeton University provides operational support, software engineering, and training for the infrastructure.Technical contributions are provided by NVIDIA, which is working to improve the performance of fusion-related computer codes, and Microsoft, which provides cloud integration through its Azure service. The platform also involves collaboration with private industry partners, including Commonwealth Fusion Systems, General Atomics, Type One Energy, and Realta Fusion. This coalition is intended to provide tools and AI models for the U.S. fusion industry to use in the development of power plant designs.
Environment Microsoft Nuclear Fusion PPPL
United States Latest News, United States Headlines
Similar News:You can also read news stories similar to this one that we have collected from other news sources.
Smart chip could slash computing energy use by up to 5,000×The new analogue in-memory chip performs computation inside memory, lowering power consumption and latency for AI and data-center workloads.
Read more »
Brace yourself: Nvidia and AMD GPU prices may rise as top AIBs adjust costsTech Product Reviews, How To, Best Ofs, deals and Advice
Read more »
Your next budget workstation GPU may be Intel Arc Pro B70Tech Product Reviews, How To, Best Ofs, deals and Advice
Read more »
GPU Price Surge: The MSRP Window Has Closed, Leaving PC Builders in a BindThe graphics card market faces significant price hikes, making it challenging to find GPUs at their recommended prices. Demand and supply issues, coupled with the AI boom, have driven up prices, impacting PC builders looking to upgrade.
Read more »
VA to slash health networks from 18 to 5 in major care system overhaulVeterans Affairs aims to streamline care by reducing regions and removing a top role, but advocates warn changes risk disrupting services for rural and high-needs veterans.
Read more »
Do You Actually Need the Best Gaming CPU?AMD’s new Ryzen 7 9850X3D seems powerful, though it's not a massive leap over what we already have.
Read more »