Job - Alert
W3-Professur „High-Performance Computing“
Deadline: 2025-05-22
Location: Germany, Hamburg
Job - Alert
W3-Professur „High-Performance Computing“
Deadline: 2025-05-22
Location: Germany, Hamburg
Jülicher Supercomputer Jupiter startet Probebetrieb
Es kommt Leben in das Jülicher Containerdorf, in dem bald der Supercomputer Jupiter in Betrieb gehen soll – als Erster in Europa mit mehr als einem ExaFLOP/s.
Updated article on Blosc2: Compute with TB-sized datasets on your own hardware, within human timeframes!
Highlights:
Outperforms NumPy by 10x ~ 100x for large computations
Maintains performance with datasets far exceeding physical memory
Integrates seamlessly with the Python data science ecosystem
Works both in-memory and on-disk with minimal performance differences
Read more: https://ironarray.io/blog/compute-bigger
JobSet: Neue API für verteilte ML- und HPC-Anwendungen auf Kubernetes
Die neue Open-Source-API JobSet soll flexiblere und vielfältigere Konfigurationsoptionen für groß angelegte verteilte HPC- und ML-Anwendungsfälle liefern.
SK Hynix and TSMC lead the HBM memory market, driven by AI and high-performance computing demands, reshaping the global semiconductor industry.The semiconductor industry is witnessing a seismic shift as High Bandwidth Memory (HBM) becomes the cornerstone of AI and high-performance computing. #AI #HBM #highperformancecomputing #memory #semiconductors #SKHynix #technology #TSMC #redrobot
https://redrobot.online/2025/03/asian-tech-giants-dominate-hbm-memory-market-as-ai-demand-surges/
https://www.maximizemarketresearch.com/market-report/hmc-hbm-market/2557/
HMC & HBM Market Trends: Growth, Challenges, and Future Outlook
SoftBank Expands AI Ambitions with $6.5B Ampere Computing Acquisition
#SoftBank #AmpereComputing #AI #Semiconductors #ChipDesign #TechAcquisition #ArtificialIntelligence #CloudComputing #HighPerformanceComputing
https://www.techi.com/softbank-acquires-ampere-chip-designer-ai-expansion/
JUPITER: Neues Supercomputermodul für europäische KI-Fabrik
Am Forschungszentrum Jülich entsteht eine europäische KI-Fabrik. Dafür bekommt der ExaFlops-Supercomputer JUPITER ein drittes Rechenmodul.
#SC24 in Atlanta, U.S., brought together global leaders, cutting-edge technology, and a shared commitment to collaboration and innovation.
Read Bigani’s report on the SC Conference and learn why he believes advancing #highperformancecomputing and #FOSS education can empower individuals and organizations in Africa and other regions in a fast-evolving tech world : https://lpi.org/4kei
RISC-V chips for a digitally autonomous EU: 240 million euros for DARE SGA1
The "Digital Autonomy with RISC-V in Europe" (DARE) project brings together 38 partner companies and research institutes to develop high performance and AI CPUs
RISC-V-Chips für eine digital autonome(re) EU: 240 Millionen Euro für DARE SGA1
Das Projekt "Digital Autonomy with RISC-V in Europe" (DARE) vereint 38 Partnerfirmen und Forschungsinstitute, die Hochleitungs- und KI-Prozessoren entwickeln.
Minnesota: The Unsung Hero of High-Performance Computing Evolution
Minnesota's pivotal role in the evolution of High-Performance Computing (HPC) is often overshadowed by its scenic lakes. From pioneering supercomputers to shaping the future of AI and quantum computin...
https://news.lavx.hu/article/minnesota-the-unsung-hero-of-high-performance-computing-evolution
Last week the majority of the DKZ.2R personnel came together for an in-person event for the very first time! Together, we worked on refining our project plan and discussed our priorities and goals for the coming months! We are excited to see what 2025 has in store! #dataliteracy #DKZ #rdm #datascience #machinelearning #highperformancecomputing #networking
Chiplet Technology: Revolutionizing Semiconductor Design and AI Computing
The Chiplet Summit in Santa Clara showcased the burgeoning trend of chiplet-based designs, poised to transform the semiconductor landscape. With insights from industry leaders, this article delves int...
AI supercomputers are consuming energy at an unprecedented scale. With 1'500 MW of power, AI megaclusters from OpenAI, DeepSeek AI, Google DeepMind, NVIDIA, and Tesla can power 1'100'000 homes, run 1'500 high-speed trains, or match half a nuclear plant’s output. The future of AI isn’t just about intelligence - it’s about #compute, #scale, and #infrastructure.
https://www.linkedin.com/pulse/ai-supercomputers-consuming-energy-unprecedented-scale-prilutskiy-fwisf/
I was hugely impressed by this podcast. The AI industry is undergoing massive changes, with trends favoring more efficient architectures, lower-cost AI models, and geopolitical competition.
Unlocking GPU Power: A Deep Dive into High-Performance SGEMM Implementations
In the race for computational supremacy, SGEMM (Single-precision GEneral Matrix Multiply) implementations are at the forefront, pushing the limits of GPU performance. This article explores cutting-edg...
Revolutionizing HPC: MLKAPS Unleashes Machine Learning for Kernel Auto-tuning
In a groundbreaking development for high-performance computing, the MLKAPS tool leverages machine learning and adaptive sampling to automate kernel tuning, significantly enhancing performance across v...
AIST and QuEra Unveil Revolutionary Quantum-HPC Cloud Platform
In a groundbreaking collaboration, AIST and QuEra have launched a quantum-high-performance computing (HPC) cloud platform that promises to redefine computational capabilities across various industries...
https://news.lavx.hu/article/aist-and-quera-unveil-revolutionary-quantum-hpc-cloud-platform
Unlocking the Secrets of NUMA: How Expensify Mastered Extreme Concurrency
In the world of high-performance computing, few concepts are as enigmatic as Non-Uniform Memory Access (NUMA). Expensify's deep dive into their NUMA balancing techniques reveals how they harness this ...