Articles tagged with Supercomputing
The Tensor Contraction Engine (TCE) is a new software that automatically generates computer code for complex molecules, reducing the workload for researchers. By generating efficient parallel programs, TCE can save time and computational resources for projects in computational chemistry and physics.
PNNL's new supercomputer is the US' fastest operational unclassified system, featuring an 11.8T HP Integrity system with Intel Itanium2 processors running Linux. The machine will enable novel studies in atmospheric chemistry, systems biology, and materials science.
The collaboration between Rice University and IBM aims to improve the understanding of carbon nanotubes' mechanical, structural, and electronic properties. By utilizing the powerful CBEN supercomputer, researchers can tackle complex quantum mysteries and develop new algorithms for theoretical molecular science.
The tele-immersion system has been upgraded to process data in real-time, thanks to the use of a far-off supercomputing center. This allows participants to feel as if they're in each other's presence, regardless of distance. The new setup uses a large bank of digital cameras and polarized glasses to create a realistic experience.
The OptIPuter project, led by Cal-(IT)2 researchers, partners with Chiaro to deploy Enstara routing platform for high-performance networking and data analysis. The platform's innovative technologies reduce infrastructure costs and enhance packet processing efficiency.
Physicist Bryan Caron and his team have achieved a world record-breaking data transfer rate, equivalent to transferring a full compact disk in under eight seconds. The team created a dedicated 'light path' spanning 12,000 km, bypassing the public Internet to establish a new transatlantic superhighway for researchers.
BioCoRE enables real-time collaboration among scientists from diverse disciplines, leveraging shared project spaces and advanced computational tools. The virtual environment bridges geographical boundaries, facilitating cross-disciplinary collaboration and accelerated research progress.
Rice Terascale Cluster will be the first university computer in Texas with a peak performance of 1 teraflop, enabling complex mathematical simulations for drug designers and biomedical researchers. The cluster will tackle increasingly complex problems in fields like bioinformatics, physics, and computer science.
The grant will double Cornell's high-performance computing capacity, making it available for business applications and research. CTC will offer consulting services and training on parallel computing and Windows-based technologies to help businesses and institutions expand their computing environment.
Researchers have discovered that protein simulations can be completed in 7 femtoseconds, three times faster than previous methods. This breakthrough enables the efficient calculation of biologically interesting processes such as protein folding.
Researchers can now perform innovative and massive new calculations, including simulating synthetic environments and analyzing genomic data. The tera-scale capability enables the creation of artificial agents with fine granularities, providing a more accurate representation of reality.
Using a powerful supercomputer, researchers simulated materials' strength and behavior, gaining insights into fracture, work hardening, and customized properties. The study enables better understanding of earthquakes and design of new materials to resist brittle fracture.
The new HP supercomputer will have an expected total peak performance of over 8.3 teraflops, 8,300 times faster than a current personal computer. Scientists will use it to study complex chemical problems in areas like life sciences and environmental research.
A Virginia Tech team has won a grant to develop mathematically rigorous decomposition theories for solving large-scale design problems on massively parallel supercomputers. The project aims to reduce communication bottlenecks and improve concurrency in the design process.
Cosmologists have developed a new simulation model that suggests the first star in the universe formed from a cloud of hydrogen and helium at least 100 times more massive than our sun. The simulations provide insights into the chemical elements produced in stars, including metals heavier than lithium.
A team of researchers discovered a new type of carbon nanotube that is approximately 40 percent stronger than other nanotubes made using the same number of atoms. This breakthrough was achieved through supercomputer simulations at the San Diego Supercomputer Center and the University of Michigan.
The National Science Foundation is awarding $53 million to build a distributed terascale facility, supporting research in storm prediction, combustion engines and biology. The project will create a unique national resource for high-performance computing, connecting researchers across the US.
A new AI system, CONSA, helps bickering computer programs resolve differences and come up with a satisfactory result. The system has been successfully tested on agent-controlled robot football teams and helicopter combat simulations.
Researchers at Penn State found that Beowulf clusters, composed of low-cost personal computers linked together, can offer supercomputer-like capabilities. The clusters can run multiple processors in parallel, providing a cost-effective solution for scientists who need to run complex codes.
Cplant Cluster, a self-made supercomputer at Sandia National Laboratories, solves benchmark problems at 232.6 billion operations per second with its 580-node assembly. The system's scalability and ability to work together make it useful for complex applications.
The Chiba City Project aims to advance highly scalable open source software development using state-of-the-art Linux clusters. The 512-CPU Linux cluster will be opened to the U.S. research community for collaborative development.
A cluster of 256 Intel Pentium III microprocessors linked together has created the largest
The Max Planck Institute for Gravitational Physics has released the public beta of the Cactus Computational Toolkit 4.0, a collaborative environment for solving partial differential equations in physics and engineering. The toolkit allows individual scientists to plug their own computing applications into a modular framework, enabling ...
The Modular-MPC 1/16 system by Aspex Microsystems Ltd. offers high-speed 3D data visualization with 10 frames-per-second performance for 16 million data elements. It achieves over 25 billion operations per second and is suitable for complex applications like computational fluid dynamics.
UC Engineers create RACE, a technique that redesigns computer hardware to adapt to user demands. The approach results in significant speed boosts for applications like image processing, enabling desktop computers to approach supercomputer speeds.
A team of scientists reached a supercomputing milestone by simulating metallic magnetism at 1.002 Teraflops on a Cray T3E supercomputer. The achievement showcases the power and capabilities of parallel computers, enabling the study of larger systems and new phenomena.
The Columbia supercomputer, QCDSP, can perform 400 billion calculations per second to simulate the three-trillion-degree conditions at the Big Bang. This will help physicists understand particle properties and calculate accurate masses and decay rates.
A team of astronomers used a supercomputer to simulate the growth of cosmic structure in a region comparable to the entire observable Universe. The simulation revealed patterns of dark matter walls and filaments, as well as massive voids and giant galaxy clusters.
The GUSTO grid will provide pervasive access to supercomputing capabilities, enabling new problem-solving techniques and distributed supercomputing. Ten groups will use Globus software and resources for various applications, including remote visualization and tele-immersion.
Sandia's achievement of the one-teraflops mark demonstrates its expertise in advanced software applications and high-fidelity 3-D simulation. This breakthrough enables scientists to simulate complex nuclear-weapon performance and predict safety against accidental nuclear explosions, marking a new era in computing.
The COMPS project aims to develop a unique networked cluster of largely off-the-shelf equipment, which will be put to work by scientists and then evaluated. The challenge is to overcome communication delays or latency that have been inherent when clusters of computers are linked together via a network.
Researchers at the Pittsburgh Supercomputing Center have developed a new brain mapping technology that can create realistic 3D images of brain activity in real-time. This breakthrough uses functional MRI data to produce high-resolution images, overcoming the previous bottleneck in processing time.
Boston University has introduced the Origin2000, a cutting-edge supercomputer designed to boost research capacity and enable scientists to tackle complex problems. The new system will be made available through the Center's MARINER Project, providing resources for other schools, universities, and businesses.
Dr. Neal Lane raises questions about dumping of Japanese vector supercomputers in the US market, affecting domestic industries, and calls for swift investigation by federal agencies like Commerce and International Trade Commission.