🎉 Emulated FP64 matrix multiplications and more have landed in NVIDIA cuBLAS. Learn how to take advantage of Tensor Core performance for algorithms that use matrix multiplication without changing your code. #HPC ✅ Faster and more power-efficient FP64 matmuls ✅ Control over mantissa bits for experimenting with accuracy and performance ✅ Support on #NVIDIABlackwell, Hopper, Ada, and Ampere architectures Read the tech blog ➡️ https://nvda.ws/4pwki38
Offer – Development Land Portfolio (NRW & Wilhelmshaven, Germany) We are pleased to present two exceptional development properties located in Germany, both suitable for large-scale energy, digital infrastructure, and industrial applications. 1) Main Property – North Rhine-Westphalia (NRW), Germany - Total land area: 382,475 m² - Location: North Rhine-Westphalia (NRW), Germany - Purchase price: EUR 76.5 million Potential Developments: - 5 Data Centers, each with 50 MW - 1 Battery Storage Facility – 121 MW/day - 1 Gas Power Plant – 6 × 63 MW modules Wilhelmshafen: 37,400 m2, EUR 14,4 mio. This site offers an outstanding opportunity for a modern, integrated, high-efficiency energy and industrial complex. H.Schemberg tradejs@t-online.de
The shift to Blackwell‑powered “AI factories” really feels like a paradigm‑change for enterprise‑scale AI and data centers. For companies in regulated fields — like health‑tech or biomedical devices — this could unlock powerful data‑driven insights and automation. From your side: which use‑cases do you think will benefit most from this infrastructure?
Thanks for sharing 👌
This is a significant advancement for computational efficiency. Leveraging Tensor Core performance without altering code is a game-changer for developers. Excited to see how this impacts HPC workflows.
Exciting development! Leveraging Tensor Core performance for FP64 matrix multiplications without changing existing code is a game-changer for HPC applications. Innovations like this not only boost computational efficiency but also open new possibilities for scientific research and advanced simulations. #HPC #NVIDIA #TensorCores #HighPerformanceComputing #Innovation #TechLeadership
Huge leap for HPC — FP64 matmuls with Tensor Core speed is game-changing.
Dive deeper into cuBLAS features and enhancements ➡️ https://nvda.ws/4aeh8wz