From egng at lbl.gov Thu Jun 3 11:26:15 2010 From: egng at lbl.gov (Esmond G. Ng) Date: Thu Jun 3 11:30:20 2010 Subject: [BANANA] Berkeley Lab - Computing Sciences Seminar - Thursday, 6/10/2010, 10:00am Message-ID: <01E72B9D-2C73-418C-B552-CDA44E80F266@lbl.gov> Berkeley Lab - Computing Sciences Seminar Date: Thursday, June 10, 2010 Time: 10:00am - 11:30am Location: Bldg. 50B, Room 2222 Speaker: Professor S.J. Ben Yoo Department of Electrical and Computer Engineering University of California, Davis Title: Intelligent Optical Networking and Computing for Future Cyberinfrastructure Abstract: Future networks will benefit from agile and intelligent reconfigurations of high-capacity traffic in the core. We will discuss optical-label switching (OLS) networks capable of seamlessly integrating packet, circuit, and flow traffic with scalability to multi-petabit-per-second capacity. Communications between the distributed agents and the network management are facilitated by the labels and the statistical summary of the labels provide information for intelligent network management and control. We will further discuss the UC Davis OLS networking testbed with OLS core routers and OLS edge routers. The prototyped OLS core routers support optical packets, bursts, and circuits with full interoperability, and achieve rapid switching (~600 psec) with ultra-low latency (~12 ns). Field trial and testbed demonstrations including video multicast/unicast and FTP applications, 1,001 hop cascaded OLS router operation, 477 km SPRINT field fiber network trials, and possible future experiments with ESnet will also be discussed. The second part of the talk will discuss scalability and energy-efficiency of future cloud computing. We will introduce a new class of optical switching system designed for interconnecting computing systems. The optical switching system exploits a single stage wavelength-routing switch with a pipelined arbiter. The switching fabric is contention-free and scalable beyond 2 milion port x 2 million port and 42 petabit/second aggregate switching capacity in a single stage switch. Simulation results indicate ~50 ns latency with 94% normalized throughput at 0.96 load in the optical switching system. Additional Notes: S. J. Ben Yoo serves as Director of UC Davis CITRIS (Center for Information Technology Research in the Interest of Society) and Professor of Electrical Engineering at University of California at Davis (UC Davis). His research at UC Davis includes future Internet architectures, high-performance optical switching systems, nano photonic-electronic systems integration for next generation networking and computing systems. His recent demonstrations included optical label switching routers with 1000 times improvement in performance/power efficiency over conventional electronic routers. Prior to joining UC Davis in 1999, he was a Senior Research Scientist at Bellcore, leading technical efforts in optical networking research and systems integration for DARPA sponsored MONET and NGI projects. He also participated in the advanced technology demonstration network/multiwavelength optical networking (ATD/MONET) systems integration, and a number of standardization activities. Prior to joining Bellcore in 1991, he conducted research on nonlinear optical processes in quantum wells, a four-wave-mixing study of relaxation mechanisms in dye molecules, and ultrafast diffusion-driven photodetectors at Stanford University (BS?84, MS?86, PhD?91, Stanford University). Prof. Yoo serves as an Associate Editor for IEEE Photonic Technology Letters, and Guest Editor for IEEE/OSA Journal of Lightwave Technology, IEEE Journal of Special Topics in Quantum Electronics, and OSA Journal of Optical Networks. Prof. Yoo is a Fellow of IEEE, a Fellow of OSA, and a recipient of the DARPA Award for Sustained Excellence in 1997, the Bellcore CEO Award in 1998, and the Mid-Career Research Faculty Award (UC Davis) in 2004. He is a Co-Chair of Photonics in Switching 2010. Host of Seminar: Inder Monga -------------- next part -------------- An HTML attachment was scrubbed... URL: http://csmr.ca.sandia.gov/pipermail/banana/attachments/20100603/729a60c7/attachment.html From egng at lbl.gov Fri Jun 11 12:22:44 2010 From: egng at lbl.gov (Esmond G. Ng) Date: Fri Jun 11 12:27:11 2010 Subject: [BANANA] Berkeley Lab - Computing Sciences Seminar - Thursday, 6/17/2010, 2:00pm Message-ID: <4C128D04.5080500@lbl.gov> Berkeley Lab - Computing Sciences Seminar Date: Thursday, June 17, 2010 Time: 2:00pm - 3:30pm Location: Bldg. 50B, Room 4205 Speaker: Joseph Eto Energy and Environmental Technologies Division Lawrence Berkeley National Laboratory Title: Consortium for Electric Reliability Technology Solutions Organization and Research Overview Abstract: The U.S. electric power system is in the midst of a fundamental transition from a centrally planned and utility-controlled structure to one that will depend on competitive market forces for investment, operations, and reliability management. Electricity system operators are being challenged to maintain the reliability of the grid and support economic transfers of power as the industry's structure changes and market rules evolve. Meanwhile, the U.S. economy depends more than ever on reliable and high-quality electricity supplies. New technologies are needed to prevent major outages such as those experienced on August 14, 2003, which cost an estimated $4-10 billion and affected 10's of millions of customers. The Consortium for Electric Reliability Technology Solutions (CERTS) was formed in 1999 to research, develop, and disseminate new methods, tools, and technologies to protect and enhance the reliability of the U.S. electric power system and efficiency of competitive electricity market. CERTS' Vision is to: Transform the electricity grid into an intelligent network that can sense and respond automatically to changing flows of power and emerging problems; Enhance reliability management through market mechanisms, including transparency of real-time information on the status of the grid; Empower customers to manage their energy use and reliability needs in response to real-time market price signals; and Seamlessly integrate distributed technologies?including those for generation, storage, controls, and communications?to support the reliability needs of both the grid and individual customers. The founding members of CERTS are LBNL, ORNL, PNNL, SNL, NSF?s Power System Engineering Research Center, and the Electric Power Group. LBNL serves as the program office for CERTS (http://certs.lbl.gov/). Additional Notes: Joseph H. Eto is a staff scientist in the Environmental Energy Technologies Division of the Lawrence Berkeley National Laboratory. He has over 25 years of experience conducting research in the energy field. His principal responsibility is management of the Consortium for Electric Reliability Technology Solutions, which is conducting multi-institution public interest research to maintain and enhance the reliability of power systems during the transition to restructured electricity markets. Mr. Eto has served the U.S. Department of Energy on a variety of topics related to electricity reliability. In 2009 and 2006, he led coordination of technical support for preparation of the DOE National Electric Transmission Congestion Studies. In 2003, he was appointed to serve on the Electricity System Working Group of the U.S.-Canada Power Outage Task Force, which is investigating the Aug. 14 blackout. In 2001/2, he led coordination of technical support to the U.S. Department of Energy (DOE) for the preparation of the National Transmission Grid Study. Prior to this assignment, he served on the Secretary of Energy?s Power Outage Study Team, which conducted independent investigations of high-profile outages that occurred during the summer of 1999. Mr. Eto?s research currently focuses on assessing, from a national perspective, the economic value of electricity reliability, with a special emphasis on the growing importance of power quality. He has authored over 150 publications on electricity reliability, power quality, demand response, distributed generation, energy-technology market transformation, utility integrated resource planning and demand-side management, and building energy-efficiency technologies. In 1989, he received both the Crosby Field Award for best technical paper and the Willis H. Carrier Award for best presentation by an author under the age of 32 from the American Society of Heating, Refrigeration and Air-conditioning Engineers. Mr. Eto received an A.B. in philosophy of science and an M.S. in energy and resources from the University of California, Berkeley. He is a registered professional Mechanical Engineer in the State of California. Host of Seminar: Inder Monga ----- Computing Sciences Seminars Web Site: http://crd.lbl.gov/CSSeminars. For additional information, such as site access or directions to the conference room, please contact CSSeminars-Help@hpcrd.lbl.gov. Web Contact: CSSeminars-Help@hpcrd.lbl.gov From egng at lbl.gov Wed Jun 16 23:39:44 2010 From: egng at lbl.gov (Esmond G. Ng) Date: Wed Jun 16 23:44:38 2010 Subject: [BANANA] Berkeley Lab - Computing Sciences Seminar - Monday, June 21, 2010, 11:00am Message-ID: <4C19C330.2090808@lbl.gov> Berkeley Lab - Computing Sciences Seminar Date: Monday, June 21, 2010 Time: 11:00am - 12:00pm Location: Bldg. 50F, Room 1647 Speaker: Surendra Byna Systems Architecture Department NEC Laboratories America, Inc., Princeton, NJ Title: Server-Push Architecture for Data Access Performance Optimization Abstract: In the last few years, we have seen enormous changes in the fa?ade of computing. Supercomputing research is targeting exa-scale computing. Almost every computing device is being equipped with multi-core and many-core processors. Parallel computing has moved from the edge to the center of both the computer science research and IT industry due to the emergence of multi-core processors. Despite all these advances, data access delay has been a major reason for poor sustained performance of systems. The key to achieving efficiency in computing is to improve data access performance. In this talk, I will discuss our efforts in improving data access performance. Our efforts can be divided into memory performance modeling and parallel I/O performance optimization. We developed memory performance prediction models based on data access patterns that are useful to choose effective optimization and prefetching strategies, with low overhead. I present our models to predict memory access cost, classifying the memory cost from communication and middleware latency. We have utilized these models to improve the performance of message passing interface (MPI) derived data types. Our approach in optimizing parallel I/O performance is server-based data pushing. In traditional data prefetching, client has to predict what data an application will access in the future and issue prefetching requests. However, due to the prediction overhead, aggressive and accurate prediction methods have been given low priority. In our method, we separate the prediction overhead onto a dedicated server and let the server push data closer to applications. I will talk about applying this method in parallel I/O and file systems. Host of Seminar: Arie Shoshani ------- Computing Sciences Seminars Web Site: http://crd.lbl.gov/CSSeminars. For additional information, such as site access or directions to the conference room, please contact CSSeminars-Help@hpcrd.lbl.gov. Web Contact: CSSeminars-Help@hpcrd.lbl.gov From egng at lbl.gov Thu Jun 17 13:11:37 2010 From: egng at lbl.gov (Esmond G. Ng) Date: Thu Jun 17 13:18:39 2010 Subject: [BANANA] Berkeley Lab - Computing Sciences Seminar - Monday, June 21, 2010, 10:00am Message-ID: <4C1A8179.6000904@lbl.gov> Berkeley Lab - Computing Sciences Seminar Date: Monday, June 21, 2010 Time: 10:00am - 11:00am Location: Bldg. 50F, Room 1647 Speaker: Dr. Farid Harhad Louisiana State University Title: Evolving Time Surfaces in a Virtual Stirred Tank Abstract: The complexity of large scale computational fluid dynamic simulations demand powerful tools to investigate the numerical results. Time surfaces are the natural higher-dimensional extension of time lines, the evolution of a seed line of particles in the flow of a vector field. Adaptive refinement of the evolving surface is mandatory for high quality under reasonable computation times. In contrast to the lower-dimensional time line, there is a new set of refinement criteria that may trigger the refinement of a triangular initial surface, such as based on triangle degeneracy, triangle area, surface curvature etc. In this talk we describe the computation of time surfaces for initially spherical surfaces. The evolution of such virtual ?bubbles'' supports analysis of the mixing quality in a stirred tank CFD simulation. We discuss the performance of various possible refinement algorithms, how to interface alternative software solutions and how to effectively deliver the research to the end-users, involving specially designed hardware representing the algorithmic parameters. Host of Seminar: Wes Bethel ----- Computing Sciences Seminars Web Site: http://crd.lbl.gov/CSSeminars. For additional information, such as site access or directions to the conference room, please contact CSSeminars-Help@hpcrd.lbl.gov. Web Contact: CSSeminars-Help@hpcrd.lbl.gov From egng at lbl.gov Fri Jun 18 13:14:11 2010 From: egng at lbl.gov (Esmond G. Ng) Date: Fri Jun 18 13:18:57 2010 Subject: [BANANA] Berkeley Lab - Computing Sciences Seminar - Thursday, June 24, 2010, 10:00am Message-ID: <4C1BD393.80502@lbl.gov> Berkeley Lab - Computing Sciences Seminar Date: Thursday, June 24, 2010 Time: 10:00am - 11:00am Location: Bldg. 50F, Room 1647 Speaker: Tanu Malik Cyber Center and Indiana Center for Database Systems Purdue University Title: Large-Scale Data Management for the Sciences Abstract: Modern scientific repositories are growing rapidly in size. Scientists are increasingly interested in viewing the latest data as part of query results. Current scientific middleware systems, however, assume repositories are static. Thus, they cannot answer scientific queries with the latest data. The queries, instead, are routed to the repository until data at the middle-ware system is refreshed. In data-intensive scientific disciplines, such as astronomy, indiscriminate query routing or data refreshing often results in runaway network costs. This severely affects the performance and scalability of the repositories and makes poor use of the middleware system. In this talk, I will present Delta a dynamic data middleware system for rapidly-growing scientific repositories. Delta's key component is a decision framework that adaptively decouples data objects, choosing to keep some data object at the middleware, when they are heavily queried, and keeping some data objects at the repository, when they are heavily updated. Our algorithm profiles incoming workload to search for optimal data decoupling that reduces network costs. It leverages formal concepts from the network flow problem, and is robust to evolving scientific workloads. Distributed applications such as the Delta framework often rely on /a priori/ knowledge of query cardinalities to make optimization decisions. In this context, I will present a black-box approach to selectivity estimation that is suitable for distributed applications. We evaluate the efficacy of Delta, through a prototype implementation, by running query traces collected from a real astronomy survey. Additional Information: Tanu Malik is a Research Assistant Professor with the Cyber Center and Indiana Center for Database Systems at Purdue University. Tanu received her Ph.D. and M.S in Computer Science from Johns Hopkins University and her B.Tech in Civil Engineering from Indian Institute of Technology, (IIT), Kanpur. Her current research focuses on scientific data management, particularly in the areas of distributed data management, improving network and I/O performance of large data sets, data provenance, data mining techniques for analyzing scientific queries, and self-managing techniques. Host of Seminar: Arie Shoshani ----- Computing Sciences Seminars Web Site: http://crd.lbl.gov/CSSeminars. For additional information, such as site access or directions to the conference room, please contact CSSeminars-Help@hpcrd.lbl.gov. Web Contact: CSSeminars-Help@hpcrd.lbl.gov