CRD’s de Jong Named to Editorial Board of Computer Physics Communications Journal
Bert de Jong, leader of CRD’s Computational Chemistry, Materials & Climate Group, has been named as Specialist Editor for the journal Computer Physics Communications, an international journal and program library for computational physics and physical chemistry. Computer Physics Communications publishes research papers and computer program descriptions in computational physics and physical chemistry. The focus is on computational methods and techniques rather than results.
The journal’s Editorial Board consists of leaders in computational physics. De Jong notes that he is “excited to be involved with such a quality journal” that has an impact factor of over three. The journal also distinguishes itself by publishing software and maintaining an online software library.
De Jong has a background in general chemistry, chemical engineering and high performance computational chemistry, with specialization and strong capabilities in modeling the bottom of the periodic system. He is a main developer of the NWChem software, one of four developers of the unique fully relativistic software MOLFDIR for quantum chemistry, and contributed to multiple other software suites in computational chemistry.
Pacific Research Platform Awarded CENIC’s 2016 Innovations in Networking Award for Experimental Applications
The Pacific Research Platform (PRP), an NSF-funded, science-driven high-capacity data-centric “freeway system” linking universities, national labs and supercomputing centers on the West Coast, has been selected by the Corporation for Education Network Initiatives in California (CENIC) as a recipient of the 2016 Innovations in Networking Award for Experimental Applications.
ESnet is a key technical partner in the PRP, which integrates Science DMZs, developed by ESnet as secure network enclaves for high-speed, data-intensive science, thereby creating a secure, seamless fabric that will enable researchers worldwide to collaborate.
“ESnet is committed to working closely with the Pacific Research Platform to leverage the Science DMZ and Science Engagement concepts to enable collaborating scientists to advance their research,” said ESnet Network Engineer Eli Dart.
The PRP will enable fast and secure data transfers between participating campuses, which include all 10 University of California campuses, Stanford, Caltech, USC, and San Diego State University – all of which are connected via the 100 Gbps CENIC Network. The PRP extends to include the University of Washington, Montana State, the University of Hawaii System, Northwestern University, UIC, and internationally to the University of Amsterdam. Since the PRP was funded, other partners have joined, including the University of Tokyo, and Clemson University. The PRP provides high-speed links to five supercomputer centers (UCSD’s SDSC, LBNL’s NERSC, NCAR, NCSA, and NASA’s NAS) as well as the Open Science Grid and NSF’s Chameleon cloud.
The PRP includes science teams in five research areas, including Astronomy and Astrophysics Data Analysis. Peter Nugent, CRD’s deputy director for scientific engagement, co-leads the team on telescope surveys.
ESnet’s Monga Delivers Plenary Talk in New Zealand
Inder Monga, acting ESnet Director and ESnet Chief Technology Officer, gave an invited talk last month at the eResearch NZ conference in New Zealand. In his talk, Monga discussed established and emerging design patterns in networks.
Monga pointed out that all R&E networks have a shared fate. No one organization is big enough or rich enough to build a global network but scientists want to collaborate with people no matter where in the world they are, so scientific data has to cross each other’s networks all the time.
“We depend on each other. We depend on all networks being excellent,” said Monga. This means we all have to help each other, it’s no good having one excellent network if the networks it connects to aren’t up to scratch.
REANNZ, the Research and Education Advanced Network New Zealand, hosted the conference and has posted a blog about Monga’s talk, complete with a video of his presentation.
This Week’s Computing Sciences Seminars
Tuesday, March 15
Rethink the Kitchen Sink: Experimental Systems for Exascale and Any Scale
3 to 4 p.m., Wang Hall – Bldg. 59, Room 3104
Kyle C. Hale Department of Electrical Engineering and Computer Science Northeastern University
We can now build parallel systems at the scale of several thousands of machines and millions of CPU cores. The growing scale and increasing heterogeneity of these high-end machines present challenges for parallel systems that are already beginning to appear in consumer devices as well. These challenges offer a unique opportunity to rethink the systems software stack and consider systems designed expressly for parallelism and the changing hardware landscape.
In this talk I will describe two such systems. The first is the Hybrid Runtime (HRT), which is a fusion of a light-weight OS kernel framework (an Aerokernel) and a parallel runtime system. HRTs are not limited by restricted access to hardware or mismatched kernel abstractions that typically arise with the pairing of a complex runtime and a general-purpose OS like Linux. In this model, the HRT essentially is the kernel. The second system, called Multiverse, is a bridging mechanism for HRTs that allows a runtime developer to transform legacy runtime systems and applications into HRTs that run as kernels and enjoy full access to hardware.
Wednesday, March 16
Processing Astronomy Imagery Using Big Data Technology
11 a.m. to 12 p.m., Bldg. 50B, Room 4205
Zhao Zhang AMPLab & Berkeley Institute for Data Science University of California, Berkeley
Scientific analyses commonly compose multiple single-process programs into a dataflow. An end-to-end dataflow of single-process programs is known as a many-task application. Typically, HPC tools are used to parallelize these analyses. In this work, we investigate an alternate approach that uses Apache Spark—a modern platform for data intensive computing—to parallelize many-task applications. Using Apache Spark, we implement Kira, a flexible and distributed astronomy image processing toolkit. We then use the Kira toolkit to implement the Kira SE application for extracting sources from astronomy images. Using Kira SE as a case study, we study the programming flexibility, dataflow richness, scheduling capacity and performance of Apache Spark running on the EC2 cloud. By exploiting data locality, Kira SE achieves a 4.1× speedup over an equivalent C program when analyzing a 1TB dataset using 512 cores on the Amazon EC2 cloud. Furthermore, we show that by leveraging software originally designed for big data infrastructure, we are able to use the Amazon EC2 cloud to achieve a 1.8× speedup over the C implementation running on the NERSC Edison supercomputer, when holding core count constant. Using the same implementation of Kira SE, a 128-core EC2 cloud deployment that uses Spark Streaming can achieve second-scale latency with a sustained throughput of ∼600 MB/s. Our experience with Kira demonstrates that data intensive computing platforms like Apache Spark are a performant alternative for many-task scientific applications.
BSE Division Seminar @ JBEI: Organization, Dynamics and Coordination of Cortical Networks in Humans and Rodents
12pm to 1pm, JBEI Seminar Room (Lab employees show badge at door)
Kristofer Bouchard, Molecules to Minds (M2M), Berkeley Lab
How distributed neural circuits gives rise to coordinated behaviors is poorly understood. In this talk, Bouchard will describe the organization, dynamics and coordination of human sensorimotor cortex during the act of speaking. Utilizing directly analogous recordings to those in humans, he will present results of multi-scale electrophysiology and optogenetics in rodents. Time permitting, he will describe ongoing activities at LBL/DOE towards the BRAIN initiative.
Bouchard’s research addresses how distributed neural circuits gives rise to coordinated behaviors. I combine multi-scale electrophysiology and optogenetics in behaving rodents with advanced analytics to investigate functional organization and dynamic coordination in sensorimotor cortex. Bouchard develops software to analyze data, including novel machine learning algorithms and deep learning approaches. Understanding mechanisms of coordination may inform remediation strategies for developmental coordination disorder, schizophrenia, and neural prosthetics.
CITRIS Research Exchange Seminar: The Future of Travel Demand
12 to 1pm, Banatao Auditorium, Sutardja Dai Hall, UC Berkeley
Susan Handy, University of California, Davis
Auto use is growing globally, along with the environmental and societal problems it generates, climate change most notably. A recent slowdown in the growth in vehicle-miles-traveled in the U.S. and Europe, however, has raised hopes that some parts of the world may have reached “peak driving.” What is behind this recent slowdown, and is it a true or temporary peak? Answering these questions requires a close look at the behavior underlying these trends and at the numerous economic, technological, and cultural changes that are driving changes in travel behavior. In this presentation, I share findings from studies from UC Davis and elsewhere on some of the critical changes that are shaping the future of travel demand. Register to reserve a seat.
Friday, March 18
BIDS Data Science Lecture Series: Research Data Publication in the Evolving World of Scholarly Communication
1:10 to 2:30pm, 190 Doe Library, UC Berkeley
Todd Vision, Department of Biology, University of North Carolina at Chapel Hill
Two different fronts in the campaign to bring data into the fold of scholarly communication have seen remarkable progress in recent years. One is the shift in policy and practice towards greater availability of research data underlying publications. Another is the advent of many large-scale efforts to extract structured information out of the scientific literature. Scientific fields dominated by individual investigators and customized data collection pose special challenges to both of these fronts but also unique opportunities. Vision will discuss some recent findings that inform the design, potential, and limitations of systems to achieve both these ends. My aim is to stimulate a conversation about the future of data publication in “small science” that considers the interplay of both technical and cultural factors.
Link of the Week: Garbage In, Garbage Out
It’s Garbage Week at Atlas Obscura and that, of course, includes computing garbage. In two articles, the online magazine explores the origins of “Garbage In, Garbage Out,” and dives into why only Apple users “trash” files, while others recycle, shred or dump.