第137回

題目：Enhancing Scientific Data Management for Exascale
講演者：Scott Klasky (Senior Scientist ORNL)
講演要旨: As we continue toward exascale, scientific data volume is continuing to scale and becoming more burdensome to manage. In this talk, we lay out opportunities to enhance state of the art data management techniques. We emphasize well principled data compression, and using it to achieve progressive refinement. This can both accelerate I/O and afford the user increased flexibility when she interacts with the data. The formulation naturally maps onto enabling partitioning of the progressively improving-quality representations of a data quantity into different media-type destinations, to keep the highest priority information as close as possible to the computation, and take advantage of deepening memory/storage hierarchies in ways not previously possible. Careful monitoring is requisite to our vision, not only to verify that compression has not eliminated salient features in the data, but also to better understand the performance of massively parallel scientific applications. Increased mathematical rigor would be ideal, to help bring compression on a better-understood theoretical footing, closer to the relevant scientific theory, more aware of constraints imposed by the science, and more tightly error-controlled. Throughout, we highlight pathfinding research we have begun exploring these related topics, and comment toward future work that will be needed.

題目：Enhancing Scientific Data Management for Exascale (Part 1 の続き)
講演者：Scott Klasky (Senior Scientist ORNL)

題目：Creating ADIOS-2 for scientific exascale data
講演者：Scott Klasky (Senior Scientist ORNL)
講演要旨: What is Scientific Exascale data? For some it is really big data from scientific experiments and simulations. We use it in Data Intensive Science, which is an acknowledgement that as simulations and experiments continue to generate larger amounts of data, we must turn our attention on how to move, store, manage, analyze and visualize this data in a timely fashion. We are already seeing Petascale simulations produce close to 100 PB per simulation, and we are hearing simulations for Exascale computing trying to approach 100 EB of data per week. Clearly the cost of “write once read never” is becoming too expensive and we must start to create software eco-systems to help us cope with this flood of data from scientific instruments and calculations. We have built the idea of I/O staging in the Adaptable I/O system (ADIOS) to ingest, reduce, and move data on HPC systems and over the WAN to other computational resources, and my talk focuses on creating a software ecosystem which employs these techniques to cope with the extreme amounts of data being produced in the DOE. Furthermore, Exascale data must be re-purposed in time in order to validate the results against physics experiments, such as the ITER fusion tokamak. This creates new challenges which must be explored and developed into an overarching infrastructure for scientific data. Our goal is to create an I/O framework that addresses most of the use-cases arising from both the Exascale challenges and the new scientific instruments coming on-line in the next 10 years.

題目：Reflecting on the goal and baseline for exascale computing
講演者：Thomas Schulthess (Swiss CSCS/ETH Director)
講演要旨: Application performance is given much emphasis in discussions of exascale computing. A 50-fold increase in sustained performance over today’s applications running on multi-petaflops supercomputing platforms should be the expected target for exascale systems deployed early next decade. In the present talk, we will reflect on what this means in practice and how much these exascale systems will advance the state of the art. Experience with today’s platforms show that there can be an order of magnitude difference in performance within a given class of numerical methods, depending only on choice of architecture and implementation. This bears the questions on what our baseline is, over which the performance improvements of exascale systems will be measured. Furthermore, how close will these exascale systems bring us to deliver on scientific goals, such as convection resolving global climate simulations or throughput oriented computations for meteorology?