News Feature Channel: Supercomputing

New AI Model Improves Prediction Power for Genomics Related to Disease

Los Alamos National Laboratory — Fri, 08 Nov 2024 09:10:45 EST

To understand the workings of DNA in relation to disease, scientists at Los Alamos National Laboratory have developed the first multimodal deep learning model of its kind, EPBDxDNABERT-2, capable of ascertaining the precise relationship between transcription factors, proteins that regulate gene activities, leveraging an aspect of DNA called DNA breathing, in which the double-helix structure opens and closes spontaneously. The model has the potential to aid in the design of drugs used to treat diseases that originate in gene activity.

Scientists Calculate Predictions for Meson Measurements

Brookhaven National Laboratory — Wed, 06 Nov 2024 15:00:39 EST

Nuclear physics theorists have demonstrated that complex calculations run on supercomputers can accurately predict the distribution of electric charges in mesons, particles made of a quark and an antiquark. The calculations also help validate a method that will be used to make predictions for and analyze data from high-energy experiments at the future Electron-Ion Collider (EIC) at Brookhaven National Laboratory.

Argonne Team Breaks New Ground in AI-Driven Protein Design

Argonne National Laboratory — Wed, 06 Nov 2024 11:50:24 EST

An Argonne-led team developed an AI-driven multimodal framework that incorporates experimental data and text-based narratives to accelerate the design of new proteins. A finalist for the Gordon Bell Prize, the effort used five top supercomputers, including Argonne's Aurora exascale system.

Aiming to Shape a New Tech Revolution

University of Miami — Sun, 03 Nov 2024 22:20:27 EST

Engineering professor Mingzhe Chen is working on several projects that could improve on our existing wireless networks. For this work, he was recently named as the Knight Foundation chair in data science and artificial intelligence.

Frontier Users' Exascale Climate Emulator Nominated for Gordon Bell Climate Prize

Oak Ridge Leadership Computing Facility — Thu, 31 Oct 2024 17:05:57 EST

A multi-institutional team of researchers led by the King Abdullah University of Science and Technology, or KAUST, Saudi Arabia, has been nominated for the Association for Computing Machinery's 2024 Gordon Bell Prize for Climate Modelling.

Celebrating Spooky Science

Department of Energy, Office of Science — Wed, 30 Oct 2024 09:15:48 EST

Science isn't scary, but it can be spooky! For Halloween, we're highlighting some "spooky" research that the Department of Energy's (DOE) Office of Science supports. In fact, the physics community celebrates Dark Matter Day on Halloween! We hope that this roundup puts you in the mood to dig a little deeper into the wonders that surround you.

Everything You Always Wanted to Know About Large Language Models for Science (but Were Afraid to Ask)

Argonne National Laboratory — Mon, 28 Oct 2024 11:45:32 EST

Large language models are changing the way that people create and communicate, but they can also change the way we do science. Researchers from Argonne National Laboratory hope to demystify these tools and share how they're shaping future research.

Researchers Use Summit to Track Down Nuclear Fission's Elusive Scission Neutron

Oak Ridge National Laboratory — Fri, 25 Oct 2024 14:50:48 EST

Researchers from the University of Washington, Seattle, or UW, and Los Alamos National Laboratory used the Summit supercomputer at the Department of Energy's Oak Ridge National Laboratory to answer one of fission's biggest questions: What exactly happens during the nucleus's "neck rupture" as it splits in two?

Researchers Succeed in Taking 3D X-ray Images of a Skyrmion

Lawrence Berkeley National Laboratory — Tue, 22 Oct 2024 11:00:00 EST

A difficult-to-describe nanoscale object called the magnetic skyrmion might one day yield new microelectronic devices that can do much more--for example, massive data storage--all while consuming much less power. But researchers need a more detailed understanding of skyrmions if they are ever to be used reliably in computational devices, including quantum computers. Berkeley Lab scientists led a project to make 3D X-ray images of skyrmions that can characterize or measure the orientations of spins inside the whole object.

Material Stimulated by Light Pulses Could Be Leap Toward More Energy-Efficient Supercomputing

Argonne National Laboratory — Wed, 16 Oct 2024 14:40:34 EST

In an Argonne-led project, researchers used X-ray microscopy to discover a ferroelectric material that tailors its response to controlled ultrafast external stimuli, such as light pulses. The material might be applicable to energy-efficient microelectronics.

Butterfly-inspired AI technology takes flight

Penn State Materials Research Institute — Wed, 16 Oct 2024 08:20:53 EST

When it comes to mating, two things matter for Heliconius butterflies: the look and the smell of their potential partner. The black and orange butterflies have incredibly small brains, yet they must process both sensory inputs at the same time -- which is more than current artificial intelligence (AI) technologies can achieve without significant energy consumption. To make AI as smart as the butterflies, a team of Penn State researchers has created a multi-sensory AI platform that is both more advanced and uses less energy than other AI technologies.

Scientists Accelerate Uranium Beam with Record Power

Department of Energy, Office of Science — Wed, 09 Oct 2024 14:05:26 EST

Researchers at the Facility for Rare Isotope Beams reached a new milestone in isotope studies, accelerating a high-power beam of uranium ions to a record 10.4 kilowatts of continuous beam power to a target. The beam enabled scientists to produce and identify three new isotopes, gallium-88, arsenic-93, and selenium-96.

Stopping Off-the-Wall Behavior in Fusion Reactors

Princeton Plasma Physics Laboratory — Mon, 07 Oct 2024 08:30:42 EST

New experimental results suggest that sprinkling boron into a tokamak could shield the wall of the fusion vessel and prevent atoms from the wall from getting into the plasma. A new computer modeling framework shows the boron powder may only need to be sprinkled from one location. The experimental results and computer modeling framework will be presented this week at the 66th Annual Meeting of the American Physical Society Division of Plasma Physics in Atlanta.

Argonne Researchers Crack a Key Problem with Sodium-Ion Batteries for Electric Vehicles and Grid Energy Storage

Argonne National Laboratory — Thu, 26 Sep 2024 15:05:31 EST

New method for cathode preparation prevents the particle cracking that caused performance decline with cycling of sodium-ion batteries, which offer a cheaper, more abundant alternative to lithium-ion batteries.

Enhancing Hurricane Forecasts: A Game-Changer in Lessening Catastrophic Impacts

Pittsburgh Supercomputing Center — Thu, 26 Sep 2024 13:05:41 EST

A team from the University of Houston found that, when they reduced estimates of atmospheric friction of storms, their predictions on PSC's Bridges-2 improved markedly over standard storm predictions. This advancement promises better planning to lessen the effects of storms on people and possibly aid emergency storm responses.

Retiring -- and Shredding -- the Alpine Storage System

Oak Ridge Leadership Computing Facility — Thu, 19 Sep 2024 11:05:18 EST

Over the summer, crews began the Alpine storage system, shredding over 40,000 hard drives with the help of ShredPro Secure, a local East Tennessee business. This partnership not only reduced costs and sped up the process but also established a more efficient and secure method for decommissioning large-scale computing systems in the future.

Smoother Surfaces Make for Better Particle Accelerators

Department of Energy, Office of Science — Wed, 18 Sep 2024 15:05:29 EST

A new toolkit helps researchers build optimal superconducting radiofrequency (SRF) cavities that form the backbone of advanced particle accelerators. The cavities' cleanliness, shape, and roughness of their inner surfaces contribute to their efficiency. In tests of the toolkit, scientists found that smoother cavities function more efficiently.

Mapping Out Matter's Building Blocks in 3D

Thomas Jefferson National Accelerator Facility — Tue, 17 Sep 2024 15:30:00 EST

Deep inside what we perceive as solid matter, the landscape is anything but stationary. The interior of the building blocks of the atom's nucleus -- particles called hadrons that most of us would recognize as protons and neutrons -- are made up of a seething mixture of interacting quarks and gluons, known collectively as partons. The HadStruc collaboration has now come together to map out these partons and disentangle how they interact to form hadrons. Their latest findings were recently published in the Journal of High Energy Physics.

Computer Engineers at ORNL Pioneer Approaches to Energy Efficient Supercomputing

Oak Ridge National Laboratory — Mon, 16 Sep 2024 11:05:33 EST

As high-tech companies ramp up construction of massive data centers to meet the business boom in artificial intelligence, one component is becoming an increasingly rare commodity: electricity. With decades of experience in making HPC more energy efficient, the OLCF may serve as a resource for best "bang for the buck" practices in a suddenly burgeoning industry.

Argonne Training Program Immerses Attendees in Extreme-Scale Computing

Argonne National Laboratory — Thu, 12 Sep 2024 17:05:32 EST

The 12th annual Argonne Training Program on Extreme-Scale Computing (ATPESC) offers intensive two-week training for next-generation scientists, computer experts, data analysts and others aiming to infuse their computing research with new vibrancy.