SHF: Small: Expediting the Execution of Machine Learning Applications on Multi-GPU Infrastructure with Architecture Awareness and Runtime Support

SHF：小型：通过架构意识和运行时支持加快多 GPU 基础设施上机器学习应用程序的执行

• 批准号: 2154973
• 负责人: Xulong Tang
• 金额: $ 59.99万
• 依托单位: University of Pittsburgh
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2022
• 资助国家: 美国
• 起止时间: 2022-06-15 至 2025-05-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2154973&HistoricalAwards=false
• 关键词: SHF; Small; Expediting; Execution; Machine

Deep Neural Networks (DNNs) have become one of the most popular machine-learning techniques for solving real-world problems in object classification, autonomous vehicles, natural language processing, etc. Due to the ever-growing problem size and complexity, the training and inference of DNN models are increasingly time-consuming and require enormous computing resources. As such, multi-GPU infrastructure is a desirable platform that has been widely used in modern DNN tasks. However, the delivered DNN execution scalability is severely limited due to architectural unawareness and lacking easy-to-use runtime support. This research uncovers and addresses the architectural bottlenecks of DNN executions. The outcome of this research is expected to achieve scalable DNN executions on multi-GPU infrastructure. The educational and outreach components of this project include (i) new course projects on multi-GPU infrastructure integrated into graduate-level computer architecture courses; (ii) engaging undergraduate students in the research activities through senior Capstone project courses and an outreach program at PI’s institute; and (iii) increasing the participation and visibility of female and minority students in computer architecture, computer science, and engineering.This research is set to uncover and address the architectural bottlenecks of DNN executions on multi-GPUs. Specifically: 1) It identifies address translation as an essential bottleneck in multi-GPU performance. It redesigns the Translation Lookaside Buffer (TLB) hierarchy and the page table walk for both single-tenant and multi-tenant DNN executions on multi-GPU infrastructure. 2) It investigates the data-movement overheads in data parallelism and model parallelism of modern DNN applications. It proposes architecture-aware data distillation and neuron-based model partitioning to mitigate the data movement overheads. 3) It proposes a runtime framework that fosters the usage of multi-GPUs through enhanced programmability, which allows dynamic and automatic virtual kernel to physical kernel generation during execution.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.

深度神经网络（DNN）已成为解决对象分类、自动驾驶汽车、自然语言处理等现实世界问题的最流行的机器学习技术之一。由于问题规模和复杂性不断增长，DNN 模型的训练和推理越来越耗时，并且需要大量的计算资源。因此，多 GPU 基础设施是一个理想的平台，已广泛应用于现代 DNN 任务。然而，由于架构意识不足和缺乏易于使用的运行时支持，所提供的 DNN 执行可扩展性受到严重限制。这项研究揭示并解决了 DNN 执行的架构瓶颈。这项研究的成果预计将在多 GPU 基础设施上实现可扩展的 DNN 执行。该项目的教育和推广部分包括 (i) 将多 GPU 基础设施集成到研究生计算机架构课程中的新课程项目； (ii) 通过高级 Capstone 项目课程和 PI 研究所的外展计划让本科生参与研究活动； (iii) 提高女性和少数族裔学生在计算机体系结构、计算机科学和工程领域的参与度和知名度。这项研究旨在发现并解决 DNN 在多 GPU 上执行的体系结构瓶颈。具体来说： 1) 它将地址转换确定为多 GPU 性能的重要瓶颈。它重新设计了多 GPU 基础设施上单租户和多租户 DNN 执行的转换后备缓冲区 (TLB) 层次结构和页表遍历。 2) 研究现代 DNN 应用程序的数据并行性和模型并行性中的数据移动开销。它提出了架构感知数据蒸馏和基于神经元的模型分区，以减轻数据移动开销。 3) 它提出了一个运行时框架，通过增强的可编程性促进多 GPU 的使用，允许在执行过程中动态自动生成虚拟内核到物理内核。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

SmartFRZ: An Efficient Training Framework using Attention-Based Layer Freezing

• DOI: 10.48550/arxiv.2401.16720
• 发表时间: 2024-01
• 期刊: ArXiv
• 作者: Sheng Li;Geng Yuan;Yuezhen Dai;Youtao Zhang;Yanzhi Wang;Xulong Tang

Trans-FW: Short Circuiting Page Table Walk in Multi-GPU Systems via Remote Forwarding

• DOI: 10.1109/hpca56546.2023.10071054
• 发表时间: 2023-02
• 期刊: 2023 IEEE International Symposium on High-Performance Computer Architecture (HPCA)
• 作者: Bingyao Li;Jieming Yin;Anup Holey;Youtao Zhang;Jun Yang;Xulong Tang

Adversarial Prefetch: New Cross-Core Cache Side Channel Attacks

• DOI: 10.1109/sp46214.2022.9833692
• 发表时间: 2021-10
• 期刊: 2022 IEEE Symposium on Security and Privacy (SP)
• 作者: Yanan Guo;Andrew Zigerelli;Youtao Zhang;Jun Yang

CEGMA: Coordinated Elastic Graph Matching Acceleration for Graph Matching Networks

• DOI: 10.1109/hpca56546.2023.10070956
• 发表时间: 2023-02
• 期刊: 2023 IEEE International Symposium on High-Performance Computer Architecture (HPCA)
• 作者: Yuezhen Dai;Youtao Zhang;Xulong Tang

AB-ORAM: Constructing Adjustable Buckets for Space Reduction in Ring ORAM

• DOI: 10.1109/hpca56546.2023.10071064
• 发表时间: 2023-02
• 期刊: 2023 IEEE International Symposium on High-Performance Computer Architecture (HPCA)
• 作者: Mehrnoosh Raoufi;Jun Yang;Xulong Tang;Youtao Zhang