InTheLoop | 08.25.2014
New Project is the ACME of Addressing Climate Change
Eight Department of Energy national laboratories, including Lawrence Berkeley National Laboratory, are combining forces with the National Center for Atmospheric Research, four academic institutions and one private-sector company in a project to develop the most complete climate and Earth system model, yet.
The Accelerated Climate Modeling for Energy, or ACME, is designed to accelerate the development and application of fully coupled, state-of-the-science Earth system models for scientific and energy applications. The plan is to exploit advanced software and new High Performance Computing machines as they become available.
Over a planned 10-year span, the project aim is to conduct simulations and modeling on the most sophisticated HPC machines as they become available, i.e., 100-plus petaflop machines and eventually exascale supercomputers. The team initially will use U.S. Department of Energy (DOE) Office of Science Leadership Computing Facilities at Oak Ridge and Argonne national laboratories. And the model will be optimized for deployment at the Department of Energy's National Energy Research Scientific Computing Center (NERSC) located at Berkeley Lab. »Read more.
ESnet's Science DMZ Model Speeds the Flow of Data at UNM
Data from the University of New Mexico’s Cancer Center next-generation genome sequencers is now flying across a 10 Gbps link to the university’s Center for Advanced Research Computing, thanks to the Science DMZ model pioneered by ESnet, according to an article posted by UNM.
According to the article, “This point-to-point connection is a first step toward establishing a campus-wide research network at UNM. The connection is based on the “Science DMZ” model formalized by the Department of Energy’s ESnet in 2010. The new link delivers a low-latency, high-bandwidth, unfiltered connection via UNM’s campus network.” »Read more.
Juliane Müller Kicks Off Alvarez Fellowship with Friday Seminar
Juliane Müeller recently joined Berkeley Lab’s Center for Computational Sciences and Engineering (CCSE) as the 2014 Alvarez Fellow. At CCSE she will be developing optimization algorithms for computationally expensive black-box problems. There are environmental applications, such as the cleaning up of contaminated groundwater at minimal cost; alternative energy applications, like generating the maximal amount of energy from hydropower dams or kites; or calibration of climate models, which involves finding better model parameters to make predictions agree better with actual observational data,” Müller said.
These are, in fact, some problems addressed in her previous work, which she is presenting in a Computing Sciences seminar at 1pm this Friday, August 29 in room 50F-1647. »Read more.
Kate Odziomek Receives ACS Scholarship for Scientific Excellence
Katarzyna (Kate) Odziomek, a visiting graduate student in chemistry from the University of Gdansk, Poland won the Chemical Information section (CINF) Scholarship for Scientific Excellence at the 2014 American Chemical Society National Meeting & Exposition held Aug. 10-14 in San Francisco. Her winning poster presentation, “Toward Quantitative Structure Activity Relationship (QSAR) Models for Nanoparticles,” describes a quantitative algorithmic framework for extracting information about nanoparticles from scanning electron microscope (SEM) images.
Where most SEM image analyses concentrate on nanoparticle distributions, Odziomek's reveals particle shapes and surface structure. “These features are really important because the way some nanoparticles interact is heavily reliant on their morphology or surface structure,” she said.
Odziomek developed the framework over the past year while working the Visualization Group’s Daniela Ushizima and the Scientific Computing Group’s Maciej Haranczyk, both scientists from the Center for Applied Mathematics for Energy Research Applications (CAMERA). »Read more.
CITRIS Research Exchange Seminar Series Starts Sept. 3
This season's speakers include David Haussler from the Center for Biomolecular Science and Engineering, Ashok Gagdil from Lawrence Berkeley National Laboratory, and bestselling author Steven Johnson.
Free and open to the public, the CITRIS Research Exchange is a weekly dialogue highlighting leading voices on societal-scale research issues. Each hourlong seminar starts at 12pm Pacific and is hosted live in the Banatao Auditorium at Sutardja Dai Hall on the UC Berkeley campus (unless otherwise noted in the schedule). Live webcasting of each seminar is also available online and in viewing rooms at UC Davis, UC Merced, and UC Santa Cruz. Advance registration is requested. »Read more.
Applied Mathematics Seminars Series Starts Sept. 3
The Applied Mathematics Seminars Series cosponsored by Berkeley Lab and UC Berkeley kicks off September 3, the first Wednesday after the Labor Day holiday. The first speaker, Angel Rubio of Universitdad de Pais Vasco, will address "Efficient implementation of time-dependent density-functional theory to treat non-linear dynamical processes in molecular nanostructures and solids." Unless otherwise noted, all seminars are on Wednesdays from 3:30–4:30 p.m. in 939 Evans Hall on the UC Berkeley Campus. »See the full schedule.
Slim Fly: A Cost Effective Low-Diameter Network Topology
University of California, Berkeley – Computing Seminar
Monday, August 25, 3:00pm - 4:00pm, 405 Soda Hall - UC Berkeley Campus
Torsten Hoefler, ETH Zürich, Switzerland
We introduce a high-performance cost-effective network topology called Slim Fly that approaches the theoretically optimal network diameter. Slim Fly is based on graphs that approximate the solution to the degree-diameter problem. We analyze Slim Fly and compare it to both traditional and state-of-the-art networks. Our analysis shows that Slim Fly has significant advantages over other topologies in latency, bandwidth, resiliency, cost, and power consumption. Finally, we propose deadlock-free routing schemes and physical layouts for large computing centers as well as a detailed cost and power model. Slim Fly enables constructing cost effective and highly resilient datacenter and HPC networks that offer low latency and high bandwidth under different workloads such as stencil or graph computations. »More.
Friday, August 29, 1:00pm - 2:00pm, Bldg. 50F, Room 1647
Juliane Müller, 2014 Alvarez Fellow
This talk focuses on algorithms we developed for solving computationally expensive black-box global optimization problems. These problems are encountered in application areas such as, for example, structural optimization, carbon sequestration, watershed management, and climate model research where time consuming simulations have to be run in order to obtain objective (and constraint) function values. The algorithms discussed here exploit information from a repeatedly updated surrogate model (also known as response surface model or metamodel), which is a computationally inexpensive approximation of the true objective function and helps to decide at which points in the variable domain the next computationally expensive function evaluation should be done. Hence, compared to optimization algorithms that rely on numerical differentiation or evolutionary strategies, fewer expensive simulations are required to find (near) optimal solutions and the optimization time is significantly lower. We discuss algorithms for optimization problems with continuous and mixed-integer variables.