CAREER: High-Performance, Power-Aware, Distributed Computing

卡麦隆，柯克大学南卡罗来纳研究基金会智力价值：复杂现象的计算机模型加深了我们对宇宙的理解，并间接提高了我们的生活质量。大规模并行和分布式系统将被用来满足分布式仿真的计算需求。如果应用程序性能效率没有提高，使用数千个耗电组件的性能将导致无法忍受的运营成本和故障率。基础技术将是困难的。多年来，旨在提高并行和分布式系统中应用程序性能的密集研究已经导致了5%-10%的平均效率。复杂性的持续指数增长使得通过调优保持这些效率具有挑战性。大幅提高效率将需要创新。广泛的影响：完成我们的研究计划将创造出从总体上提高分布式模拟性能的技术。这影响了执行基于模拟的实验的广泛学科，包括计算物理、生物学和化学。降低大规模应用的功耗将降低计算中心的运行成本，通过减少热排放来提高系统可靠性，通过节能间接影响环境。完成我们的教育计划将提升高性能计算中的电源性能考虑因素。招聘和辅导学生将培养出具有市场技能的毕业生。将包括研究发现和工具在内的先进技术整合到南卡罗来纳大学(USC)的教育课程中，将吸引下一代计算机科学家的兴趣。该项目的完成将使国际和平研究所能够为继续为计算机科学领域作出贡献奠定坚实的基础。这项职业计划的目标也与南加州大学的机构目标一致，南加州大学是EPSCORE指定的州立大学，通过奖学金促进研究和教育。

ABSTRACTCCF- 0347683Cameron, KirkUniversity South Carolina Research Foundation Intellectual Merit: Computer models of complex phenomena deepen our understanding of the universe and indirectly improve our quality of life. Large-scale parallel and distributed systems will be used to meet the computational demands of distributed simulations. If application performance efficiencies do not improve, the performance using thousands of power hungry components will lead to intolerable operating costs and failure rates. Fundamental technologies will be difficult. Years of intense research aimed at improving the performance of applications in parallel and distributed systems have led to average efficiencies of 5-10%. The continued exponential increase in complexity makes maintaining these efficiencies through tuning challenging. Improving efficiency dramatically will require innovationBroader Impact: Completion of our research plan will create technologies that improve the performance of distributed simulations generally. This impacts a broad range of disciplines that perform simulation-based experimentation including computational physics, biology and chemistry. Reducing power consumption of large-scale applications will reduce operational costs for computational center, increase system reliability through decreased heat emissions, and impact the environment indirectly through energy conservation. Completion of our educational plan will elevate the profile of power-performance considerations in high performance computing. Recruitment and mentoring of students will produce graduates with marketable skill sets. The integration of developed technologies including research discoveries and tools into educational curriculum at the University of South Carolina (USC) will capture the interest of next generation computer scientist. Completion of this project will allow the PI to lay a solid foundation for continued contributions to the field of Computer Science. The objectives of this CAREER proposal are also in line with the institutional goals of USC, an EPSCORE designated State University, to promote research and education through scholarship.