@HPC Podcast Archives - OrionX.net - HPC News Bytes – 20260216

Episode Date: February 16, 2026

- "Ride the Wave, Build the Future: Scientific Computing in an AI World", by Dongarra, Reed, Gannon - Call for National Moonshot Program for future HPC systems - DOE Genesis Mission, 26 Challenges for National Science and Technology - NSF $100M National Quantum and Nanotechnology Infrastructure, NQNI - State of The Quantum Computing Industry - Los Alamos National Laboratory Center for Quantum Computing [audio mp3="https://orionx.net/wp-content/uploads/2026/02/HPCNB_20260216.mp3"][/audio] The post HPC News Bytes – 20260216 appeared first on OrionX.net.

Welcome to HPC Newsbytes, a weekly show about important news in the world of supercomputing, AI, and other advanced technologies. Hi, everyone. Welcome to HPC Newsbytes. I'm Doug Black of Inside HBC, and with me is Shaheen Khan of OrionX.net. There's a new article by Jack Dongara, Daniel Reed, both of whom have been guests on this podcast, by the way, along with Dennis Gannon, entitled, Ride the Wave, Build the Future, scientific computing in an AI world. This is a key issue that people in our world have been wrestling with. The article asserts that scientific computing must adapt to a landscape dominated by generative AI and hyperscale platforms rather than traditional HBC alone. It presents seven guiding

maxims. One, scientific discovery increasingly requires integrated numerical modeling and AI as peer processes. 2. Energy use and data movement, not peak flops, are the key constraints. 3. Benchmarks must reflect end-to-end hybrid workflows. 4. True system success demands end-to-end co-design. 5. Scalable prototyping with tolerance for failure is essential. 6. Curated data and trained models are strategic assets. And finally, number seven, new public-private collaboration models are needed.

They conclude with a call for a national, quote, moonshot, unquote, program aimed at achieving orders of magnitude improvements in energy efficiency through energy-aware algorithms, memory-centric architectures, and software stacks optimized for hybrid AI plus simulation workloads. I encourage our listeners to listen to episode 17 of the at HBC podcast. This was in March of 2022 when Dan Reed was a special guest to discuss a previous paper with the same co-authors. That paper was titled Reinventing High Performance Computing, challenges, and opportunities. They used the metric, quote, jewels per validated scientific outcome, unquote, which is good, and also focus on end-to-end workflows, which conjures up the old

DARPA high-productivity computing system. system HPCS project of some 20 years ago. So it's energy-centric, which translates to memory-centric and infusion of AI into simulation workflows. Formulating this as a national moonshot would signal national priority and sustain political support and formalizes it with specific goals and deadlines, similar to the XA Scale program. Seems like such a program would be highly complementary to DOE's Genesis mission, which aims to double the productivity and impact of U.S. science and engineering within a decade. It was launched

in November and proceeded to announce about two dozen agreements with the private sector. And just last week, unveiled 26 national science and technology challenges covering everything from manufacturing to critical minerals, fusion energy, and particle accelerators. The full paper is on the energy.com site. The Moonshot program in the article would be highly aligned with the Genesis mission, since it focuses on energy and efficiency metrics, co-design approach for novel architectures, and next-generation compute paradigms beyond conventional HPC. In addition to the 2022 and this 2026 paper, the authors also wrote in communications of the ACM in 2023 to discuss how semiconductor constraints and hyperscale cloud providers are shifting the center of gravity for advanced

computing. That paper is titled HPC Forecast, Cloudy and Uncertain. In a strategy we're a mindful of the Genesis mission and its shared computing resources, the National Science Foundation will invest up to $100 million in a nationwide network of open access research facilities for quantum and nanosecond technologies and workforce training. The NSF NQNI, as the program is called, will support up to 16 sites over five years for students, researchers, and industry to access fabrication and characterization tools and instrumentation. It's a good time to revisit the state of the quantum computing industry. Short answer, not much has changed. Lots of progress and lots more to be done. And of course, small progress eventually will accumulate to become a milestone, and we see a lot of that in the AI space, but not quite yet in either.

Current systems are great for experimentation, preparedness, research on quantum computing, and hybrid quantum classical pilots, but they continue to be too small and too noisy and too unstable to battle classical systems. the set of suitable applications has also not changed. Optimization, physics, and shores algorithm for prime factorization, which by itself is a huge driver of the industry. While there are some early narrow cases of useful work in quantum computing, they are too small or too simplified or run slower than classical systems or otherwise too costly. So the industry continues to be in research mode.

And if you focus specifically on quantum computing, as in a programmable and a reprogrammable system versus quantum communication or quantum sensing or a quantum computing instrument, so to say, then the industry will remain in research mode for a while and even after it breaks through. So channeling funding for quantum competing through the NSF and academic research is a great idea and wider access to quantum competing modalities will help advance the state of the art and crucially, as you mentioned, also workforce development. Another similar development with similar intent was Los Alamos National Laboratory,

consolidating its quantum computing research sites and groups under a new Center for Quantum Computing based in Los Alamos. All right, that's it for this episode. Thank you all for being with us. HPC Newsbytes is a production of Orion X in association with Inside HPC. Shaheen Khan and Doug Black host the show. Every episode is featured on Insidehpc.com and posted on Orionx.net. Thank you for listening.