Large Scale Computing Track
The Large Scale Computing track is designed to provide a variety of offerings involving large scale computing for different audiences. Sessions will look at how large scale computing is involved in generating data for users at Facebook, programming and optimizing applications for heterogeneous many-core processors using PEPPHER, delivering on urgent and complex missions in federal agencies, and combining computational science and data analysis to further our scientific understanding at research laboratories. Speakers will discuss the many possibilities presented by large scale computing, while covering the challenges and complexities. These sessions will provide insight into large scale computing activities in a variety of contexts within industry, research institutions, and government. This track is intended for anyone interested in Large Scale Computing.
The Large Scale Computing sessions on Friday, November 11th are:
- Understanding Relationships Through Data
- Addressing Heterogeneous Programming Challenges Using PEPPHER
- Incenting Innovation Across Large-Scale Technology Missions
- Computational Science Panel
Understanding Relationships Through Data
Friday, November 11th, Session 5 â€“ 10:00am â€“ 11:00am; Location: C123-124
Presenter: Ming Hua (Facebook)
Abstract: The people, the friend connections, and the objects people interact with form a dynamic social graph that Facebook creates for over 500 million users. The scale of the social graph and the variety of user generated content pose grand challenges in suggesting relevant content for users. Discuss the phenomenon observed on social networks and learn about our efforts to suggest the most relevant and interesting information to our users.
Biography:Ming Hua is a Research Scientist at Facebook, where she focuses on modeling relationships among Facebook entities and building large scale systems to improve user experience. Ming has published in premier academic journals and conferences. She serves in the Program Committees of numerous International Conferences such as SIGKDD, ICDM and PKDD. She is one of the Exhibits and Demos Co-Chairs for the 2011 IEEE International Conference on Data Mining (ICDM 2011). She was the publicity chair of the First ACM SIGKDD Workshop on Knowledge Discovery from Uncertain Data. Ming holds a Ph.D. degree in Computing Science from Simon Fraser University, Canada.
Addressing Heterogeneous Programming Challenges Using PEPPHER
Friday, November 11th, Session 6 â€“ 11:30am â€“ 12:30pm; Location: C123-124
Presenter: Beverly May Bachmayer (Intel GmbH)
Abstract: PEPPHER(www.peppher.eu) is a unified framework for programming and optimizing applications for heterogeneous many-core processors to ensure performance portability. Participants will gain an understanding of the challenges presented by highly parallel, heterogeneous architectures. The presentation will highlight how a highly collaborative group of European researchers created a holistic approach to solving these challenges. Best methods for working successfully in a group and the first demonstration from the project will be shown.
Biography: Bev Bachmayer holds a Bachelor’s degree in Computer Science from the University of Oregon (1983) and an MBA from Portland State University (1992). Bev has worked in diverse software engineering, engineering and program management positions in the US and Europe during her 28 years at Intel. Currently a technical consulting engineer, working in the HPC tools group, she is Intel’s lead on the EU PEPPHER project. Her key area of interest is performance analysis and optimization of software on new computer architectures. Additionally, Bev supports increasing the number of professional females entering computer science/engineering programs worldwide through multiple local projects.
Incenting Innovation Across Large-Scale Technology Missions
Friday, November 11th, Session 7 â€“ 3:45pm – 4:45pm; Location: C123-124
Moderator: Donna Roy (Department of Homeland Security)
Panelists: Emma Garrison-Alexander (Transportation Security Administration), Margie Graves (Department of Homeland Security), Leslie Hope (U.S. Citizenship and Immigration Services), and Sandra Peavy (Federal Law Enforcement Training Center)
Abstract: The top challenges within the government IT today mimic those found in private industry; compounded by the complexities associated with operating as a federal agency. These extraordinary senior leaders attract and maintain highly skilled, effective and motivated workforces to deliver on the urgent and complex missions in a chaotic environment of this relatively new federal Department. In the words of Grace Hopper, these dynamic leaders â€śmanage things â€“ and lead peopleâ€ť.
Ms. Roy joined the DHS in December of 2006 and current serves as the CIOâ€™s Executive for Information Sharing She is driving the development of the DHS information sharing environment including technical policy development and coordination with the White House National Security Staff and the Program Manager for the Information Sharing Environment.
Ms. Roy served as the VP for Product Development in a Fortune 200 company in the financial services sector and the VP for a major data operations division. She has over 29 years of experience culminating a data-oriented view for technology implementation that increases operational efficiency.
Dr. Garrison-Alexander current leads over $400 million of IT initiatives.
Dr. Garrison-Alexander has 20 years experience with National Security Agency, starting as an Electronic Engineer growing into leadership positions in Technology and Systems, Signals Intelligence, and Information Assurance. Her experience includes command and control of time-sensitive signals intelligence and information assurance missions. She graduated from NSAâ€™s Senior Leadership Development Program and was a member of the Defense Intelligence Senior Executive Service.
Dr. Garrison-Alexander holds a Bachelor of Science in Electrical Engineering, a Master of Science in Telecommunications Management and a Doctorate in Management with a focus on technology and information systems.
Margaret Graves is the Department of Homeland Securityâ€™s Deputy Chief Information Officer. She oversees an IT portfolio of $6.4 billion in programs. She manages operations in the functional areas of IT Portfolio Management, Applied Technology, Enterprise Architecture, Data Management, IT Security, Infrastructure Operations and Enterprise Services.
Ms. Graves also has 20 years experience in management consulting where she held executive positions and performed consulting engagements for clients. She has experience in systems engineering, business process reengineering, strategic planning, mergers and acquisitions and venture capital planning.
Ms. Graves holds a M.B.A. and a B.S. in Chemistry from the University of Virginia.
Ms. Hope was one of the founding technology participants to the inception of DHS in 2002. As Deputy CIO for USCIS, Ms. Hope is responsible for IT development, maintenance, and enhancement of the organizationâ€™s worldwide operations. USCIS is currently in the process of consolidating data centers and a transformation program moving from paper based to digital benefits processing.
With a 33-year professional career in information technology, Ms. Hope was responsible for IT operations of the former INS. Prior to her government service she held programming and management positions in the banking and airline industries
Sandy Peavy was named the FLETCâ€™s first CIO and Assistant Director in November 1999. She is responsible for providing technology for the training of our nationâ€™s law enforcement personnel. As one of the original CIOs for DHS, she helped stand up the Departmentâ€™s core communications, computing, and enterprise service, and the CIO Council structure.
Ms. Peavy has over 30 years experience in information technology with the federal government. She began her career as a GS-3 computer assistant for OPM, served 13 years as a civilian in the Air Force, helped DISA consolidate their Defense Megacenters, and served on Treasuryâ€™s CIO Council.
Computational Science Panel
Friday, November 11th, Session 8 â€“ 5:15pm – 6:15pm; Location: C123-124
Presenters: Carol S Woodward (Lawrence Livermore National Laboratory) and Janine C Bennett (Sandia National Laboratories)
Abstract: The increasing capability of large-scale high performance computers has redefined the possibilities of scientific exploration and discovery. New HPC capabilities are becoming more accessible to scientists and engineers across academia, industry, and government in domains ranging from biology to astronomy to economics to social science. The systems allow researchers to explore an extreme range of time scales (from femtoseconds to millennia) and spatial scales (from subatomic to the cosmos). As awe-inspiring as the possibilities are, the increasing complexity in both scientific models and data sets poses a significant challenge. A distinguished panel will discuss this challenge and present recent examples of how computational science and data analysis have joined forces to further our scientific understanding.