Today at Berkeley Lab

Researchers Elected as Fellows of the American Academy of Arts & Sciences

James Demmel of the Computational Research Division, and Dean Toste and Birgitta Whaley of the Chemical Sciences Division, have been elected as fellows of the American Academy of Arts and Sciences. Founded in 1780, the Academy honors exceptional scholars, leaders, artists, and innovators. More>

You can leave a comment. View the Comment Policy.

Researchers Develop Platform for Hosting Science Data Analytics Competitions

The National Nuclear Security Administration is hosting a competition to find innovative algorithms to detect non-natural radiation sources in urban environments. They’ve teamed up with researchers in the Lab’s Computational Research, Nuclear Science, and Information Technology divisions to build a Kaggle-inspired data analytics competition platform to host it. More>

You can leave a comment. View the Comment Policy.

Juliane Mueller’s Career Path a Blueprint for Other Researchers

Finding a career can be an exciting and sometimes challenging time in one’s life. But it comes naturally to Juliane Mueller of the Computational Research Division. Her rock climbing hobby is reflective of how she mapped out her career path. The first step rock climbers take is to search for a spot where they can gain a hold that will lead them to their objective. More>

You can leave a comment. View the Comment Policy.

Kepler Solves Mystery of Fast and Furious Explosions

NASA’s Kepler Space Telescope has been used to catch notoriously short-lived FELTs — Fast-Evolving Luminous Transients — in the act and determine their nature. This has allowed astronomers to quickly arrive at this model for explaining FELTs. The Lab’s David Khatami is among those contributing to the modeling effort. More>

You can leave a comment. View the Comment Policy.

El Cerrito High School Students Get Hands-On Experience During Visit

Students from the IT Academy at El Cerrito High School recently took part in a special Big Data Study Trip hosted at the Lab and sponsored by Workforce Development and Education, EESA, NERSC, and CRD. They plotted characteristics and mapped patterns using water-quality data, and brainstormed on ways to collect and analyze data from San Pablo Bay. More>

You can leave a comment. View the Comment Policy.

Researchers Discuss Challenges of Storing and Retrieving Exascale Data

Future exascale supercomputers will be able to process a quintillion calculations per second. How can users store and retrieve data quickly enough on such a system? In this podcast, CRD’s Suren Byna and NERSC’s Quincey Koziol discuss prepping HDF5 – one of the most parallel I/O libraries for high-end supercomputing systems – for exascale use.

You can leave a comment. View the Comment Policy.

Lab Researchers Aid Speedier Scientific Data Flows

Eli Dart (ESnet), Michael Wehner (CRD), and Prabhat (NERSC) recently teamed up to document the data transfer workflow, data transfer performance, and other aspects of moving approximately 56 terabytes of climate data from the distributed Coupled Model Intercomparison Project (CMIP5) archive to NERSC. More>

You can leave a comment. View the Comment Policy.

Tool Turns Large-Scale Scientific Array Data Analysis Into a Cakewalk

Computational Research Division researchers helped develop a novel scalable framework that performs scientific analysis of vast data arrays more efficiently and tens to thousands of times faster than current big data management systems. Recently, Laser Interferometer Gravitational-Wave Observatory astronomers used the tool to discover two colliding neutron stars. More>

You can leave a comment. View the Comment Policy.

Improving the Running, Scheduling of Workflows on HPC Systems

Scientists use high performance computing (HPC) to carry out scientific workflows. But the steps and programs needed to execute these workflows require time-consuming manual tasks and don’t make the most efficient use of the system. Newly released software allows HPC scheduling systems to automatically address these issues. More>

You can leave a comment. View the Comment Policy.

GraphBLAS: Building Blocks for High Performance Graph Analytics

GraphBLAS, a collaboration between researchers in academia, industry, and national research laboratories — including Aydın Buluç of the Computational Research Division — has made publicly available a collection of standardized building blocks for graph algorithms in the language of linear algebra. These will aid development in artificial intelligence, big data, and data analytics. More>

You can leave a comment. View the Comment Policy.