Scientific Application Requirements for Leadership Computing at the Exascale

The Department of Energy s Leadership Computing Facility, located at Oak Ridge National Laboratory s National Center for Computational Sciences, recently polled scientific teams that had large allocations at the center in 2007, asking them to identify computational science requirements for future exascale systems (capable of an exaflop, or 1018 floating point operations per second). These requirements are necessarily speculative, since an exascale system will not be realized until the 2015 2020 timeframe, and are expressed where possible relative to a recent petascale requirements analysis of similar science applications [1]. Our initial findings, which beg further data collection, validation, and analysis, did in fact align with many of our expectations and existing petascale requirements, yet they also contained some surprises, complete with new challenges and opportunities. First and foremost, the breadth and depth of science prospects and benefits on an exascale computing system are striking. Without a doubt, they justify a large investment, even with its inherent risks. The possibilities for return on investment (by any measure) are too large to let us ignore this opportunity. The software opportunities and challenges are enormous. In fact, as one notable computational scientist put it, the scale of questions being asked at the exascale is tremendous and the hardware has gotten way ahead of the software. We are in grave danger of failing because of a software crisis unless concerted investments and coordinating activities are undertaken to reduce and close this hardwaresoftware gap over the next decade. Key to success will be a rigorous requirement for natural mapping of algorithms to hardware in a way that complements (rather than competes with) compilers and runtime systems. The level of abstraction must be raised, and more attention must be paid to functionalities and capabilities that incorporate intent into data structures, are aware of memory hierarchy, possess fault tolerance, exploit asynchronism, and are power-consumption aware. On the other hand, we must also provide application scientists with the ability to develop software without having to become experts in the computer science components. Numerical algorithms are scattered broadly across science domains, with no one particular algorithm being ubiquitous and no one algorithm going unused. Structured grids and dense linear algebra continue to dominate, but other algorithm categories will become more common. A significant increase is projected for Monte Carlo algorithms, unstructured grids, sparse linear algebra, and particle methods, and a relative decrease foreseen in fast Fourier transforms. These projections reflect the expectation of much higher architecture concurrency and the resulting need for very high scalability. The new algorithm categories that application scientists expect to be increasingly important in the next decade include adaptive mesh refinement, implicit nonlinear systems, data assimilation, agent-based methods, parameter continuation, and optimization. The attributes of leadership computing systems expected to increase most in priority over the next decade are (in order of importance) interconnect bandwidth, memory bandwidth, mean time to interrupt, memory latency, and interconnect latency. The attributes expected to decrease most in relative priority are disk latency, archival storage capacity, disk bandwidth, wide area network bandwidth, and local storage capacity. These choices by application developers reflect the expected needs of applications or the expected reality of available hardware. One interpretation is that the increasing priorities reflect the desire to increase computational efficiency to take advantage of increasing peak flops [floating point operations per second], while the decreasing priorities reflect the expectation that computational efficiency will not increase. Per-core requirements appear to be relatively static, while aggregate requirements will grow with the system. This projection is consistent with a relatively small increase in performance per core with a dramatic increase in the number of cores. Leadership system software must face and overcome issues that will undoubtedly be exacerbated at the exascale. The operating system (OS) must be as unobtrusive as possible and possess more stability, reliability, and fault tolerance during application execution. As applications will be more likely at the exascale to experience loss of resources during an execution, the OS must mitigate such a loss with a range of responses. New fault tolerance paradigms must be developed and integrated into applications. Just as application input and output must not be an afterthought in hardware design, job management, too, must not be an afterthought in system software design. Efficient scheduling of those resources will be a major obstacle faced by leadership computing centers at the exas..