SHF: Small: Automatic, adaptive and massive parallel data processing on GPU/RDMA clusters in both synchronous and asynchronous modes

SHF：小型：在同步和异步模式下在 GPU/RDMA 集群上自动、自适应和大规模并行数据处理

• 批准号: 2005884
• 负责人: Xiaodong Zhang
• 金额: $ 45万
• 依托单位: Ohio State University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2020
• 资助国家: 美国
• 起止时间: 2020-07-01 至 2024-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2005884&HistoricalAwards=false
• 关键词: SHF; Small; Automatic; adaptive; massive

The computing ecosystem in both hardware and software is in a critical transition time, coming from several technology crises and inevitable trends. First, the continued performance improvement in general-purpose processors is no longer realistic. Second, conventional processors are increasingly inefficient in both performance and power consumption for various data-intensive applications. Finally, the deep software stack that has been developed for several decades, from instruction-set architecture all the way to the programming layer in the existing ecosystem, has added cumbersome processing and even unnecessary overhead in computing. To address the above-mentioned issues, this project remedies the computing ecosystem in an accelerator-based way. GPU (Graphic Processing Unit) and RDMA (Remote Data Memory Access) are the two external hardware accelerators considered in the project. It aims to turn efficient asynchronous computing into a reality on clusters of hardware accelerators of GPU and RDMA adaptively and automatically by removing three technical barriers in the existing ecosystem: (1) the programming-model barrier, (2) the hardware abstraction barrier, and (3) the automation barrier. The project strives to make broad and transformational impact. It is expected to influence the data-processing research community with new algorithms and effective systems implementation, and influence industries to improve their production systems in their daily computing operations serving society. The developed algorithms, source code and measurements are available online for a public and wide usage, benefiting both industrial and academic researchers. The research training to both undergraduate and graduate students address the concerns of lacking hardware-acceleration and data-analytics professionals in information technology and computing industries. The curriculum development introduces related research results to classrooms and the outreach activities encourage high school students to be interested in computing related college education. The existing computing environment does not provide programming models for asynchronous execution. It is even harder for asynchronous programming on GPU/RDMA clusters. The execution-model difference between CPUs and GPUs makes the system lack a common hardware abstraction for GPU computing and for RDMA communication and management. Asynchronous programming is hard, and an automatic tool to ensure its correctness and efficiency is highly desirable. This research project bridges the gap between asynchronous computing and GPU/RDMA. It develops an autonomous memory pool (AMP) interfacing GPU/RDMA clusters, where an intermediate representation is proposed to abstract the GPU execution and AMP constructed by an RDMA. A set of intermediate representations are developed to support asynchronous programming, so that users can easily express asynchronous computing in programming. In addition, an intermediate representation is developed to allow conventional synchronous programming to become automated asynchronous execution code. The system is tested and evaluated using representative data-processing workloads on large GPU/RDMA clusters.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.

由于多次技术危机和不可避免的趋势，硬件和软件的计算生态系统正处于关键转型时期。首先，通用处理器的持续性能改进已不再现实。其次，对于各种数据密集型应用，传统处理器的性能和功耗效率越来越低。最后，现有生态系统中从指令集架构一直到编程层，发展了数十年的深层软件堆栈增加了繁琐的处理，甚至不必要的计算开销。为了解决上述问题，该项目通过基于加速器的方式来修复计算生态系统。 GPU（图形处理单元）和RDMA（远程数据内存访问）是该项目中考虑的两个外部硬件加速器。它旨在通过消除现有生态系统中的三个技术障碍：（1）编程模型障碍，（2）硬件抽象障碍，以及自适应地、自动地将高效的异步计算在GPU和RDMA硬件加速器集群上变成现实。 (3)自动化障碍。该项目致力于产生广泛的变革性影响。预计它将通过新算法和有效的系统实施来影响数据处理研究界，并影响各行业在服务社会的日常计算操作中改进其生产系统。开发的算法、源代码和测量结果可在线供公众广泛使用，使工业和学术研究人员受益。针对本科生和研究生的研究培训解决了信息技术和计算行业缺乏硬件加速和数据分析专业人员的问题。课程开发将相关研究成果引入课堂，外展活动鼓励高中生对计算机相关大学教育产生兴趣。现有的计算环境不提供异步执行的编程模型。 GPU/RDMA集群上的异步编程就更难了。 CPU 和 GPU 之间的执行模型差异使得系统缺乏用于 GPU 计算以及 RDMA 通信和管理的通用硬件抽象。异步编程是困难的，非常需要一个自动工具来确保其正确性和效率。该研究项目弥补了异步计算和 GPU/RDMA 之间的差距。它开发了一个连接 GPU/RDMA 集群的自主内存池 (AMP)，其中提出了一种中间表示来抽象 GPU 执行和由 RDMA 构建的 AMP。开发了一组中间表示来支持异步编程，以便用户可以轻松地在编程中表达异步计算。此外，还开发了一种中间表示形式，以允许传统的同步编程变成自动化的异步执行代码。该系统使用大型 GPU/RDMA 集群上的代表性数据处理工作负载进行测试和评估。该奖项反映了 NSF 的法定使命，并通过使用基金会的智力优点和更广泛的影响审查标准进行评估，被认为值得支持。

项目成果

Automating Incremental and Asynchronous Evaluation for Recursive Aggregate Data Processing

• DOI: 10.1145/3318464.3389712
• 发表时间: 2020-05
• 期刊: Proceedings of the 2020 ACM SIGMOD International Conference on Management of Data
• 作者: Qiange Wang;Yanfeng Zhang;Hao Wang;Liang Geng;Rubao Lee;Xiaodong Zhang;Ge Yu

DBSpinner: Making a Case for Iterative Processing in Databases

• DOI: 10.1109/icde51399.2021.00273
• 发表时间: 2021-04
• 期刊: 2021 IEEE 37th International Conference on Data Engineering (ICDE)
• 作者: Sofoklis Floratos;A. Ghazal;Jason Sun;Jianjun Chen;Xiaodong Zhang

Mixer: Efficiently Understanding and Retrieving Visual Content at Web-Scale

• DOI: 10.14778/3476311.3476371
• 发表时间: 2021-07
• 期刊: Proc. VLDB Endow.
• 作者: Mengbai Xiao;An Qin;Yongwei Wu;Xinjie Huang;Xiaodong Zhang

Maze: A Cost-Efficient Video Deduplication System at Web-scale

• DOI: 10.1145/3503161.3548145
• 发表时间: 2022-10
• 期刊: Proceedings of the 30th ACM International Conference on Multimedia
• 作者: An Qin;Mengbai Xiao;Ben Huang;Xiaodong Zhang

NestGPU: Nested Query Processing on GPU

• DOI: 10.1109/icde51399.2021.00092
• 发表时间: 2021-04
• 期刊: 2021 IEEE 37th International Conference on Data Engineering (ICDE)
• 作者: Sofoklis Floratos;Mengbai Xiao;Hao Wang-;Chengxin Guo;Yuan Yuan-Yuan;Rubao Lee;Xiaodong Zhang