EAGER: Tensor500: A Streaming Analytics High Performance Computing Benchmark

EAGER：Tensor500：流分析高性能计算基准

• 批准号: 1849463
• 负责人: Tim Andersen
• 金额: $ 9.99万
• 依托单位: Boise State University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2018
• 资助国家: 美国
• 起止时间: 2018-09-01 至 2021-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1849463&HistoricalAwards=false
• 关键词: EAGER; Tensor500; Streaming; Analytics; Performance

Large-scale data analytics is essential to the National Strategic Computing Initiative (NSCI) and NSF's scientific discovery mission. Data-intensive supercomputing applications are increasingly important for handling high-performance computing workloads. However, current benchmarks and performance metrics do not provide useful information on suitability of computing systems for data-intensive applications. This project will develop a new benchmark that can be used to assess the performance of new system implementations aimed at this important class of applications. The research effort will produce a new Stream500 benchmark, complementary to Graph500, that replicates streaming analytics workload. The current benchmark capabilities will be expanded by enabling the exiting web-based infrastructure to accept submissions for multiple benchmarks.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.

大规模数据分析对于国家战略计算计划（NSCI）和NSF的科学发现任务至关重要。数据密集型超级计算应用程序对于处理高性能计算工作负载越来越重要。但是，当前的基准和性能指标并未提供有关计算系统适用于数据密集型应用程序的有用信息。 该项目将开发一个新的基准测试，该基准可用于评估针对此重要类别应用程序的新系统实施的性能。研究工作将产生新的流500基准，该基准互补，以复制流分析工作负载。当前的基准功能将通过使退出的基于Web的基础架构能够接受多个基准的提交来扩展。该奖项反映了NSF的法定任务，并被认为是通过基金会的智力优点和更广泛影响的审查标准通过评估来通过评估来获得支持的。

项目成果

A parallel sparse tensor benchmark suite on CPUs and GPUs

• DOI: 10.1145/3332466.3374513
• 发表时间: 2020-01
• 期刊: Proceedings of the 25th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming
• 作者: Jiajia Li;M. Lakshminarasimhan;Xiaolong Wu;Ang Li;C. Olschanowsky;K. Barker