TPC26: Toward Scientific AI Platforms at HPC Facilities AI is creating a new set of demands for HPC centers. Researchers are no longer... A year ago, AWS’s Thierry Pellegrino estimated that quantum computing was still four to five years away from broad commercial relevance. Speaking at TPC26 last week, however, he suggested the timeline may be accelerating. “I think 2027 is going to see a lot of advancements,” Pellegrino said, arguing that some quantum computing modalities are making progress toward the logical qubit counts needed to tackle meaningful scientific problems. The prospect of quantum computing becoming useful sooner than expected formed a central theme of Pellegrino’s keynote, which examined how advances in quantum computing, artificial intelligence and high performance computing are reshaping scientific discovery. Pellegrino also argued that researchers increasingly need access to both cloud and on-premises computing resources, depending on the scale, urgency and nature of their workloads. The result, he said, is a more flexible computing environment in which advanced computing capabilities are no longer limited to organizations that can build and operate dedicated supercomputers. “There used to be a time that a lot of us remember where building a supercomputer would take years,” Pellegrino said. Today, he added, researchers can access supercomputing resources “with the click of a button,” enabling a level of flexibility and scale that was previously available only to a handful of large national laboratories. Against that backdrop, Pellegrino outlined four areas where he believes quantum computing is most likely to deliver practical value first: physics and chemistry, cryptography, materials science, and optimization. In physics and chemistry, quantum computers could help researchers simulate molecules and other quantum systems that are difficult to model accurately using conventional computers. Materials science represents a related opportunity, with researchers exploring how quantum systems might be used to engineer
TPC26: Pellegrino Says 2027 Could Mark a <b>Quantum</b> Turning Point
Read the original article
hpcwire.com →