Discover how @allen_ai, @Cirrascale, and NVIDIA are creating a transparent, reproducible, and accessible #AI ecosystem through the OMAI project in this fireside chat. Explore the future of AI in academia, Cirrascale’s platform innovations, and real-world breakthroughs in healthcare, climate science, and beyond. 🗓 Tuesday, November 18 | ⏰ 1p.m.–2 p.m. ET Don’t miss the live demo and Q&A 👉 Register Now: nvda.ws/4nyXFd5

Discover how the intersection of supercomputing and science is being redefined. 🔥Join our NVIDIA #SC25 Fireside Chat to hear directly from the minds shaping the future of #HPC. Our own Ian Buck will lead a discussion with global leaders: Ceren Susut (DOE), Dan Stanzione (TACC), Thomas Lippert (Jülich Supercomputing Centre), Satoshi Matsuoka (RIKEN) Hear powerful, unscripted insights into how AI is revolutionizing areas like drug discovery, astrophysics, and materials science. 📅 Monday, Nov 17, 4:00 – 5:30 p.m. CT (Doors open at 3:30) 📌 Reserve your spot: nvda.ws/4oMiNh2

With CUDA 13.0 update 2, emulation of FP64 matrix multiplication was made available in cuBLAS, providing faster and more power-efficient GEMMs, guardrails for accuracy via ADP, and the ability to experiment with mantissa bit control. Details/links here: developer.nvidia.com/blog/un…

Did you know that sparse linear algebra is as applicable to AI as it is to scientific computing? Of course you did. What you might not have known is that there is an opportunity for the right individual to lead, mentor, and grow a team at NVIDIA that is focused on sparse linear algebra libraries. And, if you have the qualifications seen on the linked page, you might even want to apply for the role. nvidia.wd5.myworkdayjobs.com…

I couldn't agree more with the sentiment. Some of the more senior graduate students at OSU referred to Prof. Cull as "the oracle" (before "The Matrix") because if you were stuck on a problem, he seemed to come up with a solution as you described the issue to him. Nice guy too.

Join us for the first NVIDIA Quantum Developer Day at #GTC25! Explore the future of #QuantumComputing with talks from leading experts and NVIDIA partners. From machine learning to chemistry, discover how #quantum is transforming industries. ➡️nvda.ws/3XmUKdf

📣 @linuxfoundation's Post-Quantum Cryptography Alliance now integrates NVIDIA cuPQC, offering GPU-accelerated, quantum-resistant cryptographic protocols. #QuantumComputing #PQCA Learn more about this impactful collaboration now. nvda.ws/40SXZLD

Level up your Python performance with Warp 1.5.0! 🚀 New tile-based programming primitives for accelerated matrix multiplication and Fourier transforms. Check it out nvda.ws/3BEsICk #Python #AI #machinelearning

📣 Introducing NVIDIA cuPQC for GPU-accelerated post-quantum #cryptography. 🔒 Seamlessly transition to secure PQC protocols and protect against future #quantum threats with this easy, flexible #SDK. Learn more ➡️ nvda.ws/4f2LHEe 👩‍💻

Announcing cuPyNumeric, the #opensource distributed #acceleratedcomputing library designed to be a drop-in replacement of NumPy, enabling scientists and researchers to harness GPU acceleration at cluster scale without modifying their Python code. #SC24 ➡️ nvda.ws/493mmJe

I didn't get to talk about this as much as I wanted at #SC24 😆 I've been particularly fascinated by the possibility of emulated FP64. Take a look at this paper! arxiv.org/abs/2411.12090

Learn how researchers from @MIT, Mines Paris-PSL, and @argonne used the NVIDIA cuDSS library to solve power grid optimization problems in real-time at scales large enough to address the largest US power grids. ➡️ nvda.ws/40Xs3X6 #SC24

Jack Dongarra, John Gunnels, Harun Bayraktar, Azzam Haidar, Dan Ernst: Hardware Trends Impacting Floating-Point Computations In Scientific Applications arxiv.org/abs/2411.12090 arxiv.org/pdf/2411.12090

Great talks by Bronis de Spinski on El Captain, Rio Yokota on the Ozaki Method (which could be the future way of computing high precision DGEMM), and Erich Strohmaier on Top500 trends.

📣 Announcing cuPyNumeric, the #opensource distributed #acceleratedcomputing library designed to be a drop-in replacement of NumPy, enabling you to harness GPU acceleration at cluster scale without modifying your #Python code. ➡️ blogs.nvidia.com/blog/cupynu… #SC24

🎉 Just released: NVIDIA HPC SDK v24.11 🌟 👀 Check out the enhancements to the math libraries and improvements for C++ programming. 📥 Download: nvda.ws/48Pc0fH #HPC #HPCSDK

#SC24 is approaching - Join us November 17-22 at our partner booths on the show floor and check out a variety of virtual content throughout the week here: nvda.ws/48yY1e0 #HPC #Supercomputing

We are back at #SC24 with our BoF 📣 🚨 Mixed Feelings about Mixed Precisions 🚨 Wed, Nov 20th 12:15pm-1:15pm Room B211 w/ @HartwigAnzt, H. Bayraktar @nvidia, and @luszczekhpc #HPC #MxP hicma.kaust.edu.sa/news/mixe… The logscale does not convey a sense of AI fatigue. 🫣 👇

Identical to my first car, modulo the bird on the hood (there was a flaw in the original paint on mine, so the hood was repainted and the owner opted out replacing the "screaming eagle"). Highly enjoyable car. The wind effect with the top removed was ... memorable.

cuQuantum SDK: A toolkit for accelerating quantum simulators via High Performance GPU computing: Recent developments by Dmitry Lyakh from @nvidia