InTheLoop | 03.05.2007
The weekly newsletter for Berkeley Lab Computing Sciences
March 5, 2007
NERSC’s John Shalf Featured in Article on Challenges of Parallelism
HPCwire recently published a Q&A with John Shalf, head of NERSC’s Science Driven System Architecture Team, and UC Berkeley professor David Patterson, about parallelism in high-performance computing. The two are co-authors of a white paper called “The Landscape of Parallel Computing Research: A View from Berkeley,” which looks at the shift to parallelism in hardware and software in both the high-performance computing world and the general computer market. It identifies the challenges in designing systems that will continue to push the performance curve and offers recommendations on developing programs that will run efficiently on computers with many processor cores. Kathy Yelick, head of CRD’s Future Technologies Group, and Parry Husbands of the Scientific Computing Group were also co-authors of the paper. Read the Q&A at http://www.hpcwire.com/hpc/1288079.html. The white paper can be found at http://www.eecs.berkeley.edu/Pubs/TechRpts/2006/EECS-2006-183.pdf.
Head of German Computing Center at Juelich to Visit NERSC on March 6
Thomas Lippert, director of the Central Institute for Applied Mathematics at Research Center Juelich in Germany, will give a talk at 9:30 a.m. Tuesday, March 6, as part of a daylong visit to NERSC. Lippert’s talk, “HPC at the Research Center Juelich: Experience with BG/L and Future Plans,” will be given in 943-238 and presented via videoconference to 50B-2222.
Here’s the abstract for his presentation:
Simulation Sciences have become central to research at Forschungszentrum Juelich (FZJ) and are considered as a most important strategic asset. FZJ’s major research fields — energy, health environment and information technology — will strengthen and concentrate their already strong simulation component within the Juelich Platform for Simulation Sciences (JPSS), which is complemented by the new German Research School for Simulation Sciences.
The Juelich supercomputing centre, the John von Neumann Institute for Computing (NIC), is a national user facility providing supercomputers for the Helmholtz Association of German Research Centres and about 250 research groups at German and European universities. Juelich is a candidate for one of the four planned European supercomputer centres.
The NIC is operating the largest supercomputer in Germany since the end of 2003. It is an early adopter of the highly scalable Blue Gene technology. The NIC follows a dual supercomputing concept, with a few large simulation projects suited for highly scalable machines and requiring CPU times measured in units of teraflop/s-years running on the Blue Gene systems, and about 200 projects that need large local memory or SMP functionality running on a 10 teraflop/s IBM Power cluster. As a founding member of the European supercomputer network DEISA, Juelich is a central part of the European multi-site GPFS environment and provides the Grid workflow software UNICORE designed for transparent cooperation in a distributed supercomputer environment.
In my presentation I will give an overview of Juelich’s HPC activities and I will outline the ongoing European process to create a petascale supercomputer infrastructure. I will report on our efforts to improve the scaling of applications on the 50 teraflop/s Blue Gene/L system and I will sketch NIC’s upgrade plans as far as petaflop hardware and multi-petabyte storage are concerned.
NVIDIA Staff to Give March 9 Talk on CUDA GPU
Interested Lab employees are invited to a seminar at 11:15 a.m. Friday, March 9, during which Andy Keane and David Kirk of NVIDIA will present an overview of NVIDIA’s CUDA (Compute Unified Device Architecture) GPU. The session will be held in 50A-5132. Keane is general manager, GPU Computing Business, and Kirk is chief scientist at NVIDIA. According to NVIDIA’s Web site,
CUDA technology is a fundamentally new computing architecture that enables the GPU to solve complex computational problems in consumer, business, and technical applications. CUDA technology gives computationally intensive applications access to the tremendous processing power of NVIDIA graphics processing units (GPUs) through a revolutionary new programming interface. Providing orders of magnitude more performance and simplifying software development by using the standard C language, CUDA technology enables developers to create innovative solutions for data-intensive problems. For advanced research and language development, CUDA includes a low level assembly language layer and driver interface.
Read more at http://developer.nvidia.com/object/cuda.html.
The NVIDIA presentation will be followed by a brown bag seminar featuring two presentations on related projects at UC Berkeley. At noon, Vasily Volkov and Jim Demmel will discuss “GPUs for Dense Linear Algebra,” and at 12:30 p.m., Sara McMains will talk about “Experience Programming GPUs.” Interested staff are invited to bring their lunch and listen to the talks.
Peter Nugent to Discuss Hardware, Software, Workflow and the Fate of the Universe
On Friday, March 9, Peter Nugent of the Scientific Computing Group will give a talk on “Hardware, Software, Workflow and the Fate of the Universe” starting at noon in Room 238 at the Oakland Scientific Facility.
Nugent is a member of the SciDAC Computational Astrophysics Consortium, the Supernova Factory, the SNAP Collaboration, the GOSH Collaboration and a former member of the Supernova Cosmology Project, which was named Science Magazine’s 1998 “Breakthrough of the Year.” He has been a NERSC user since 1996. Nugent will talk about his recent work using supernovae to measure the expansion of the universe, describe his particular high performance computing challenges and solutions, and present his workflow and data analysis implementations and needs.
2007 Grace Hopper Celebration of Women in Computing Issues Call for Participation
Submissions are now being accepted for the 7th Grace Hopper Celebration of Women in Computing conference, to be held Oct. 17-20 in Orlando, Fla. This year’s theme, “I Invent the Future,” recognizes and celebrates the significant role women have in the creation, use and leadership of computing and technology.
As the world’s largest gathering of women in computing, the Grace Hopper Celebration is a technical conference featuring well-known keynote speakers and invited technical speakers, panels, workshops, new investigator technical papers, Ph.D. forums, technical posters, birds-of-a-feather sessions (BoFs), and the ACM Student Research Competition and Awards Celebration. Submissions are sought for:
- Ph.D. Forum
- Panels, Workshops, and Presentations
- Technical Posters
- ACM Student Research Competition (SRC)
- Birds-of-a-Feather Sessions
- New Investigator Technical Papers
The deadline for all submissions is Saturday, March 31. More information can be found at http://gracehopper.org/2007/participate.html.
About Computing Sciences at Berkeley Lab
The Lawrence Berkeley National Laboratory (Berkeley Lab) Computing Sciences organization provides the computing and networking resources and expertise critical to advancing the Department of Energy's research missions: developing new energy sources, improving energy efficiency, developing new materials and increasing our understanding of ourselves, our world and our universe.
ESnet, the Energy Sciences Network, provides the high-bandwidth, reliable connections that link scientists at 40 DOE research sites to each other and to experimental facilities and supercomputing centers around the country. The National Energy Research Scientific Computing Center (NERSC) powers the discoveries of 7,000-plus scientists at national laboratories and universities, including those at Berkeley Lab's Computational Research Division (CRD). CRD conducts research and development in mathematical modeling and simulation, algorithm design, data storage, management and analysis, computer system architecture and high-performance software implementation. NERSC and ESnet are Department of Energy Office of Science User Facilities.
Lawrence Berkeley National Laboratory addresses the world's most urgent scientific challenges by advancing sustainable energy, protecting human health, creating new materials, and revealing the origin and fate of the universe. Founded in 1931, Berkeley Lab's scientific expertise has been recognized with 13 Nobel prizes. The University of California manages Berkeley Lab for the DOE’s Office of Science.
DOE’s Office of Science is the single largest supporter of basic research in the physical sciences in the United States, and is working to address some of the most pressing challenges of our time. For more information, please visit science.energy.gov.