Collaborative Research: FMitF: Track I: DeepSmith: Scheduling with Quality Guarantees for Efficient DNN Model Execution

合作研究：FMitF：第一轨：DeepSmith：为高效 DNN 模型执行提供质量保证的调度

• 批准号: 2019336
• 负责人: Cunxi Yu
• 金额: $ 36.79万
• 依托单位: University of Utah
• 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: 2020-10-01 至 2024-02-29
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2019336&HistoricalAwards=false
• 关键词: Collaborative; Research; FMitF; Track; DeepSmith

The recent unprecedented growth of deep learning has led to rapid advances in a multitude of cutting-edge technologies such as computer vision, language translation, autonomous driving, and financial-fraud detection. However, realistic deep-learning models based on deep neural networks (DNNs) typically have substantial computational and memory requirements, which greatly limit their training and deployment in resource-constrained settings. The proposed research aims to employ formal methods to significantly improve the performance of DNN execution while providing useful quality guarantees that will enable a wider deployment of deep learning. This project will produce open-source software and conference tutorials to facilitate technology transfer and fruitful industry-academia interactions in a multidisciplinary community. This project proposes DeepSmith, a scheduling framework for efficient DNN model execution based on satisfiability modulo theories (SMT). The core of the proposed project includes a novel resource-constrained scheduling formulation with combined theories using SMT to exactly encode a rich set of performance and resource constraints, and a collection of advanced domain-specific SMT-solving algorithms. Moreover, a domain-specific programming language will be developed to enable the rapid development of exact scheduling using SMT and high code reusability. The resulting DeepSmith framework will allow productive exploration and deployment of SMT in DNN execution and potentially other optimization tasks in high-performance computing and hardware acceleration.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.

深度学习的最新前所未有的增长导致了许多尖端技术，例如计算机视觉，语言翻译，自动驾驶和财务欺诈检测的迅速发展。但是，基于深神经网络（DNN）的现实深度学习模型通常具有实质性的计算和内存要求，这极大地限制了他们在资源受限的设置中的培训和部署。拟议的研究旨在采用正式方法来显着提高DNN执行的性能，同时提供有用的质量保证，从而可以更广泛地部署深度学习。该项目将生产开源软件和会议教程，以促进多学科社区中的技术转移和富有成果的行业 - 阿卡迪亚互动。该项目提出了DeepSmith，这是一个基于满意度模型理论（SMT）的有效DNN模型执行的调度框架。拟议项目的核心包括一种新型的资源约束计划公式，并使用SMT结合理论来精确编码一组丰富的性能和资源约束，并集合了高级域特异性SMT解决算法。此外，将开发特定于领域的编程语言，以便能够使用SMT和高码可重复使用性快速开发精确的调度。由此产生的Deepsmith框架将允许SMT在DNN执行中的生产性探索和部署，以及在高性能计算和硬件加速中的潜在其他优化任务。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和更广泛影响的评估来评估的支持。

项目成果

Artificial Intelligence Accelerators Based on Graphene Optoelectronic Devices

• DOI: 10.1002/adpr.202100048
• 发表时间: 2021-06-01
• 期刊: ADVANCED PHOTONICS RESEARCH
• 作者: Gao, Weilu;Yu, Cunxi;Chen, Ruiyang
• 通讯作者: Chen, Ruiyang

IMpress: Large Integer Multiplication Expression Rewriting for FPGA HLS

• DOI: 10.1109/fccm53951.2022.9786123
• 发表时间: 2022-05
• 期刊: 2022 IEEE 30th Annual International Symposium on Field-Programmable Custom Computing Machines (FCCM)
• 作者: Ecenur Ustun;Ismail San;Jiaqi Yin;Cunxi Yu;Zhiru Zhang

Read your Circuit: Leveraging Word Embedding to Guide Logic Optimization

• DOI: 10.1145/3394885.3431560
• 发表时间: 2021-01
• 期刊: 2021 26th Asia and South Pacific Design Automation Conference (ASP-DAC)
• 作者: Walter Lau Neto;Matheus T. Moreira;L. Amarù;Cunxi Yu;P. Gaillardon

Exact Memory- and Communication-aware Scheduling of DNNs on Pipelined Edge TPUs

• DOI: 10.1109/sec54971.2022.00023
• 发表时间: 2022-12
• 期刊: 2022 IEEE/ACM 7th Symposium on Edge Computing (SEC)
• 作者: Jiaqi Yin;Zhiru Zhang;Cunxi Yu

Device‐System End‐to‐End Design of Photonic Neuromorphic Processor Using Reinforcement Learning

使用强化学习的光子神经形态处理器的设备-系统端-端设计

• DOI: 10.1002/lpor.202200381
• 发表时间: 2022
• 期刊: Laser & Photonics Reviews
• 影响因子: 11
• 作者: Tang, Yingheng;Zamani, Princess Tara;Chen, Ruiyang;Ma, Jianzhu;Qi, Minghao;Yu, Cunxi;Gao, Weilu
• 通讯作者: Gao, Weilu