EuroHPC Joint Undertaking (EuroHPC JU)’s Post

The Next Big Thing in AI Computing - CRAM : A team of researchers in the University of Minnesota College of Science and Engineering have demonstrated a state-of-the-art hardware device that could reduce energy consumption for artificial intelligent (AI) computing applications by a factor of at least 1,000. ** Introducing Computational Random-Access Memory (CRAM) Computational Random-Access Memory (CRAM) - a revolutionary model where data never leaves the memory. • For the first time, researchers have demonstrated CRAM's ability to process data entirely within the memory array, eliminating the need for it to leave the storage grid. The research is published in npj Unconventional Computing, a peer-reviewed scientific journal published by Nature. The researchers have multiple patents on the technology used in the device. **Key Features of CRAM: • Revolutionizes computing by integrating computation directly into memory. • Eliminates the traditional bottleneck between processing and memory in von Neumann systems. • Allows computations to occur anywhere within the memory array. • Adapts to various AI algorithms. • Reduces the need for slow, energy-intensive data transfers. Reference: https://lnkd.in/g3rSn5YD Stay tuned for more updates on this revolutionary technology and its impact on the future of AI. #AI #Innovation #Research #Technology #CRAM #UniversityofMinnesota #Sustainability #Computing #EnergyEfficiency #TechBreakthrough #ArtificialIntelligence #AcademicResearch

𝐘𝐚𝐥𝐞 𝐭𝐨 𝐁𝐮𝐢𝐥𝐝 𝐚 𝐇𝐲𝐛𝐫𝐢𝐝 𝐐𝐮𝐚𝐧𝐭𝐮𝐦-𝐂𝐥𝐚𝐬𝐬𝐢𝐜𝐚𝐥 𝐆𝐏𝐓 𝐟𝐨𝐫 𝐍𝐨𝐯𝐞𝐥 𝐌𝐨𝐥𝐞𝐜𝐮𝐥𝐞 𝐆𝐞𝐧𝐞𝐫𝐚𝐭𝐢𝐨𝐧 𝐔𝐬𝐢𝐧𝐠 𝐍𝐕𝐈𝐃𝐈𝐀 𝐂𝐔𝐃𝐀-𝐐 𝐓𝐋𝐃𝐑: Yale researchers are pushing the boundaries of quantum machine learning with a hybrid quantum-classical GPT for molecule generation, leveraging NVIDIA CUDA-Q to accelerate drug discovery. ***𝘍𝘪𝘯𝘥 𝘵𝘩𝘦 𝘧𝘶𝘭𝘭 𝘯𝘦𝘸𝘴 𝘢𝘳𝘵𝘪𝘤𝘭𝘦 𝘢𝘵𝘵𝘢𝘤𝘩𝘦𝘥 𝘸𝘪𝘵𝘩 𝘵𝘩𝘪𝘴 𝘱𝘰𝘴𝘵.*** 𝑹𝒆𝒍𝒂𝒕𝒆𝒅 𝑪𝒐𝒏𝒕𝒆𝒏𝒕 - NVIDIA's CUDA-Q platform: https://lnkd.in/er4vh7fQ #Yale #NVIDIA #AI #DrugDiscovery #QuantumComputing #MachineLearning

After decades of researching and developing parallel-computing systems—which form the bedrock of today’s AI capabilities—Professor Torsten Hoefler is ready for bigger challenges. In this exclusive interview for Supercomputing Asia, the newly appointed Chief Architect for Machine Learning at the Swiss National Supercomputing Centre (CSCS) talks about bolstering HPC systems to support more AI-driven applications, the dynamic interplay between the two fields, and the future of AI. “Deep learning is enabled by the availability of massively parallel devices that can train massively parallel models. Today, the workload of an AI system is extremely parallel, allowing it to be distributed across thousands, or even millions, of processing components,” he said. Get the full story through the link in the comments below ⬇️ — #NSCC #supercomputing #HPC #science #AI #deeplearning #AIregulation #parallelcomputing #supercomputers

Great to show with live demos how to develop AI-infused app with Cloud-based workbench. Solving real world (eg. healthcare) challenges with data is the key to deliver true values.

Bets in AI deep tech The goal of this project - assessing propositions in AI that could produce notable results in the medium term. I am drawing on in-depth reading of the scientific literature. I make public what I can. The first three propositions under the microscope: - structure and sufficiency of data - wetware - neuromorphic hardware https://lnkd.in/dsMFAWdk

Next-Gen AI Hardware Uses Spin Waves to Unlock Unmatched Computing Power 🌐💡📊 https://lnkd.in/g256ipzb #AI #SpinWaves #ReservoirComputing #IonGating #AdvancedScience #Neuromorphic #Innovation #EnergyEfficiency #MagneticMaterials #Computing

🌟 Revolutionizing Scientific Research #DeSci is transforming research by tackling traditional challenges like high computational costs. #NuNet, an open-source orchestration system, enables decentralized, peer-to-peer compute networks, giving researchers affordable access to, Data processing, Simulations, Complex analyses. Accessible computing is driving innovation and accelerating discovery. #AI #DeSci #Research

This news couldn't be more timely! 😍 Recent research has demonstrated that the IF-COMP method significantly enhances AI prediction accuracy and reliability. I’m excited to share that my #zlnn (Liquid Neural Network library) already includes an experimental implementation of IF-COMP. This feature, along with others such as Dynamic Neural Network structure, selectable backends (for hardware acceleration on GPUs and FPGAs), and memory safety (courtesy of Rust and extensive testing/SAST/DAST tools), offers users top-notch algorithms to design their own LLMs, CNNs, and other popular models. As a reminder: the zLNN library will be available to the academic community under the GNU AFFERO GENERAL PUBLIC LICENSE and to enterprises under a commercial license. This ensures that everyone can benefit from my work while the work remains fully sponsored, guaranteeing ongoing support and continuous improvements. Contact me here for details, licenses and more info: https://lnkd.in/exr-Adbm Link to the Research announcement here: https://lnkd.in/eCsE7i4W Link to IF-COMP Research paper here: https://lnkd.in/eienauSq #AI #ArtificialIntelligence, #MachineLearning, #DeepLearning, #NeuralNetworks, #AIPrediction, #LiquidNeuralNetworks, #IFCOMP, #DynamicNeuralNetworks, #HardwareAcceleration, #GPUs, #FPGAs, #RustProgramming, #Rust, #MemorySafety, #OpenSource, #AcademicResearch, #EnterpriseSolutions, #AIResearch, #TechInnovation, #SoftwareDevelopment, #AICommunity

Intel's Aurora Supercomputer Ushers in a New Era of AI-Fueled Scientific Discovery Thrilled to share the news of Intel's Aurora supercomputer, exceeding the exascale threshold at 1.012 exaflops and claiming the title of the world's fastest AI-focused system! This powerhouse isn't just about speed, it's about innovation. Designed for AI from the ground up, Aurora empowers researchers to tackle complex challenges like: ✓Mapping the human brain's 80 billion neurons ✓Revolutionizing drug discovery with machine learning ✓Unraveling the mysteries of high-energy physics with deep learning This is just the beginning! The future of scientific discovery is bright thanks to Aurora's ability to process massive datasets and generate groundbreaking insights. #AI #Supercomputing #ScientificDiscovery #Intel

AMAX is proud to have deployed the first NVIDIA DGX™ system for Universidad del Bío Bío in Chile, enhancing the university’s AI research capabilities and supporting advancements across various disciplines. The system, named 'Pewma' by the university, draws from a Mapuche term meaning dreams and vision, symbolizing their mission to drive innovative research and address complex challenges. As a global leader in high-performance computing, AMAX provides advanced solutions to academic institutions worldwide. 🌍 From North America to South America, we’re enabling research that drives progress. 🔗 Read the full Press Release Here: https://lnkd.in/e9MSG7WM #DGXElite #AI #Academia #AMAXEngineering