 |
 |
|||||||
 |
||||||||
 |
 |
 |
 |
 |
 |
 |
 |
|
 |
|
|
|
|
|
|
|
|
About CHECS
In 2003, some of us led the design and construction of System X, a 2200 processor supercomputer which debuted as the 3rd fastest system, with the best price/performance, in the world. This facility is providing unprecedented raw computational power to scientists and engineers, and now gives us a unique opportunity to design a new generation of high-end computing systems and environments.
The focus of CHECS is on computer science systems research. Center members recognize that high-end resources must be powerful in a broad sense (i.e., high-performance, high-capacity, high-throughput, high-reliability, etc.), and at the same time they must be more usable and affordable than current HPC systems. Toward that end, the Center is pursuing a broad research agenda in areas such as processor and memory architectures, operating systems, run-time systems, communication subsystems, fault-tolerance, scheduling and load-balancing, power-aware systems and algorithms, numerical algorithms and programming models. The goal is to build computing systems and environments that can efficiently and usably span the scales from department-sized machines to national-scale resources. Our approach is forward-looking, practical and collaborative. The strategy is to harness trends in computing technology and market forces to build real systems that will meet the day-to-day needs of computational scientists. Our work on computing environments is based on real job mixes, made up of real codes, incorporating feedback from real CSE users. With the lessons learned from developing proven, well-tested systems, the Center will lead the design and deployment of the next generation of high-end systems. The Center works closely with collaborators from across the COE as well as from other colleges. It supplies computing cycles, algorithmic expertise, training, consulting, partnerships in pursuing funding, opportunities for students to work in interdisciplinary teams, and an opportunity for the CSE community to share best-practices. In turn, CSE collaborators motivate and inform CHECS research, and supply all-important application suites on which to test newly developed systems.
Motivation
The advent of powerful supercomputers has enabled computational science and engineering to join the ranks of theory and experiment as a means of understanding the world around us. Computational science studies the universe through first principles analysis, enabling us to observe systems that are too small, too large or too dangerous for direct experimental observation. While high-end computing systems continue to improve in peak performance, success stories in computational science and engineering (CSE) are still few and far between. Achieving the full potential of CSE will require stronger collaboration between computer scientists and CSE practitioners. CSE and CS researchers must push each other and push their respective fields. Better results require a full-court press: better modeling, better algorithms, better software and better hardware. Leading CSE research teams rarely succeed without significant advances in all of these categories. Working in isolation, CSE practitioners must wait for advanced technology to trickle down from other places, and CS researchers miss a diverse and demanding set of motivating applications, as well as the chance to contribute to problems of tremendous significance to society.In 2003, some of us led the design and construction of System X, a 2200 processor supercomputer which debuted as the 3rd fastest system, with the best price/performance, in the world. This facility is providing unprecedented raw computational power to scientists and engineers, and now gives us a unique opportunity to design a new generation of high-end computing systems and environments.
Mission
The mission of the Center for High-End Computing Systems (CHECS) is world-class computer systems research in the service of high-end computing. CHECS members investigate a broad array of problems and design a wide range of technologies-all with the goal of developing the next generation of powerful and usable high-end computing resources. Established in September 2005, and supported by Virginia Tech's College of Engineering (COE), CHECS depends critically on collaboration with COE computational science and engineering practitioners. CHECS provides computational and consulting services to CSE researchers; in return CSE researchers supply challenging problems which serve as motivation and evaluation testbeds for new systems research.The focus of CHECS is on computer science systems research. Center members recognize that high-end resources must be powerful in a broad sense (i.e., high-performance, high-capacity, high-throughput, high-reliability, etc.), and at the same time they must be more usable and affordable than current HPC systems. Toward that end, the Center is pursuing a broad research agenda in areas such as processor and memory architectures, operating systems, run-time systems, communication subsystems, fault-tolerance, scheduling and load-balancing, power-aware systems and algorithms, numerical algorithms and programming models. The goal is to build computing systems and environments that can efficiently and usably span the scales from department-sized machines to national-scale resources. Our approach is forward-looking, practical and collaborative. The strategy is to harness trends in computing technology and market forces to build real systems that will meet the day-to-day needs of computational scientists. Our work on computing environments is based on real job mixes, made up of real codes, incorporating feedback from real CSE users. With the lessons learned from developing proven, well-tested systems, the Center will lead the design and deployment of the next generation of high-end systems. The Center works closely with collaborators from across the COE as well as from other colleges. It supplies computing cycles, algorithmic expertise, training, consulting, partnerships in pursuing funding, opportunities for students to work in interdisciplinary teams, and an opportunity for the CSE community to share best-practices. In turn, CSE collaborators motivate and inform CHECS research, and supply all-important application suites on which to test newly developed systems.
© 2006 Virginia Tech