Please visit the CLUSTER website regularly for updates, since there may be schedule revisions.
September 9, Wednesday
Exascale System Architecture Directions: (Rebalancing of Memory with Compute)
Speaker: Alan Gara (Intel)
Abstract: New memory technologies are changing the way we architect systems and open up the possibility to dramatically address the bandwidth challenges we have in today’s system. To exploit this exciting opportunity a refactoring of the memory+compute appears to be a promising direction.
It is a familiar story that we need to address power. Possible non-evolutionary architectural directions and concepts will be discussed that may enable us to realize better performance/Watt enabling more performance within a reasonable power envelope. While there are hardware-only approaches to address this, additional options will also be discussed that are opened up when we allow for new software directions together with hardware architecture innovations.
Short biography: Alan Gara is Intel’s chief architect for exascale computing. One aspect of this is to lead the Coral technical effort.
As exascale architect, he is leading a team of Intel architects in pathfinding the future for Xeon-Phi compute and Omni-Scale fabric directions on the approach to exascale.
Prior to joining Intel he was an IBM fellow and Chief Architect for three generations of the Blue Gene platform, which achieved the #1 position in the Top500 list multiple times. He also led the team responsible for the BlueGene supercomputer receiving a national medal of technology and innovation in 2009.
Alan received two Gordon Bell prizes (1998 and 2006) and the Seymour Cray award in 2010. He has over 70 publications in computer science and physics and more than 130 US patents in the area of computer design and architecture.
Alan received his PhD in physics from the University of Wisconsin, Madison.
September 10, Thursday
The Road to Extreme Scale and the End of Moore’s Law
Speaker: Marc Snir, Argonne National Laboratory
Abstract: Since the late 80’es, we have leveraged the fast advance in microprocessor performance in order to build ever more powerful supercomputers; this fast advance was largely due to the exponential growth in the density of integrated circuits, as predicted by Moore’s Law. As Moore’s Law approaches its limits, it has become more difficult to continue pushing the performance of supercomputers: Clock speed has stopped increasing a decade ago, and power constraints are leading toward the use of simpler cores; as a result, increases in performance levels have required a more than proportional increase in concurrency levels.
Moore’s Law faces increasing obstacles in the coming decade, as we approach the ultimate limit of atomic sizes: Circuit miniaturization is increasingly expensive, while the benefits from such a miniaturization are decreasing; and different technologies, such as memory and cpu, are progressing at different speeds, requiring ever more complex architectures.
I shall discuss in the talk the obstacles faced by the semiconductor industry and their impact on extreme scale computing; I shall outline the likely evolution toward exascale performance in the next decade and speculate on the evolution of supercomputing beyond exascale.
Short biography: Marc Snir is Director of the Mathematics and Computer Science Division at the Argonne National Laboratory and Michael Faiman and Saburo Muroga Professor in the Department of Computer Science at the University of Illinois at Urbana-Champaign. He currently pursues research in parallel computing.
He was head of the Computer Science Department at Illinois from 2001 to 2007. Until 2001 he was a senior manager at the IBM T. J. Watson Research Center where he led the Scalable Parallel Systems research group that was responsible for major contributions to the IBM SP scalable parallel system and to the IBM Blue Gene system.
Marc Snir received a Ph.D. in Mathematics from the Hebrew University of Jerusalem in 1979, worked at NYU on the NYU Ultracomputer project in 1980-1982, and was at the Hebrew University of Jerusalem in 1982-1986, before joining IBM. Marc Snir was a major contributor to the design of the Message Passing Interface. He has published numerous papers and given many presentations on computational complexity, parallel algorithms, parallel architectures, interconnection networks, parallel languages and libraries and parallel programming environments.
Marc is Argonne Distinguished Fellow, AAAS Fellow, ACM Fellow and IEEE Fellow. He has Erdos number 2 and is a mathematical descendant of Jacques Salomon Hadamard. He recently won the IEEE Award for Excellence in Scalable Computing and the IEEE Seymour Cray Computer Engineering Award.
September 11, Friday
A Journey to Exascale Computing
Speaker: William J. Harrod (DOE ASCR)
Abstract: Exascale computing is a shared international pursuit aimed at creating a new class of high performance computing systems that can achieve a hundred times the sustained performance of today’s petascale computers while limiting growth in space and power requirements. Although the primary goal of this pursuit is to develop leading edge computing assets for new scientific discovery, medical science, climate modeling, and other compute- and data-intensive applications, the resulting technologies will have a profound impact on all future computing systems down to laptops and handheld devices.
Computing is now at a critical crossroads. We can no longer proceed down the path of steady but incremental progress to which we have become accustomed. Thus, exascale computing is not simply an effort to provide the next level of computational power by creative scaling up of current petascale computing systems.
New architectures will be required to achieve the exascale computing goals. Although there are many daunting challenges, which have been identified and extensively examined in numerous previous studies, past and ongoing pilot projects have indicated the feasibility of achieving the exascale goals. However, development of exascale technology is not just a hardware problem. A significant investment in system software, programming environments, and applications algorithms and codes is required as well.
This presentation will focus on the strategy and plans for developing and deploying energy efficient, highly programmable exascale computers by the early 2020s. Various challenges will be discussed, including, technical, programmatic, and policy issues.
Short biography: Dr. William Harrod is currently the Research Division Director in the Advanced Scientific Computing Research program in the Office of Science at the Department of Energy (DOE). Dr Harrod received his PhD in Mathematics from the University of Tennessee at Knoxville. He has served as a Principal Engineer and an Engineering Manager of Computational and Applied Mathematics at SGI and the Section leader of the Mathematical Software Group at Cray.
Before joining DOE, Dr. Harrod was a Program Manager in the Information Processing Technology Office (IPTO) in the Defense Advanced Research Projects Agency (DARPA). While at DARPA he was responsible for a wide range of research activities related to the development and exploitation of advanced computing technologies and applications in support of the DoD and the Military Services. In this role, he led two internationally recognized hallmark studies that set the computer performance goals and strategy for achieving extreme scale computing. He also initiated a significant new research effort in Ubiquitous High Performance Computing (UHPS).