SHF: Small: A Hybrid NVM based Computing Architecture for Machine Learning Applications

如今，企业越来越多地把目光投向内部，将大量未处理或丢弃的数据视为有待挖掘的资源。因此，一个被称为内存密集型计算的新兴领域已经引起了工业界和学术界的兴趣，主要是由各种新兴的非易失性存储技术（NVMs）推动的。机器学习（ML）应用程序正成为内存密集型计算的目标，利用ML应用程序的独特属性，以数量级提高其分布式性能。因此，在这种内存密集型计算环境中执行的机器学习应用程序非常需要高效、灵活和可扩展。ML应用程序从磁盘驱动器处理大量数据，由于磁盘访问延迟而增加延迟。该项目探索和设计了新技术，使机器学习应用程序充分利用nvm中中间数据的持久性的好处，这大大减少了磁盘I/ o，从而减少了数据处理时间。该项目还涉及课程开发，并为妇女、少数民族和代表性不足的学生进入研究和研究生课程提供更多途径。本项目专注于内存密集型计算系统的一个开放挑战，即如何为机器学习应用提供高效率、低成本和更大的灵活性，特别是在异构环境下。本项目提出了一种基于混合NVM的计算架构，具有有效的数据共享和通信策略，以优化ML应用程序的文件管理，资源分配和数据通信。本研究围绕两个关键设计展开：1)基于由字节和块可寻址设备组成的混合NVM池的文件和数据管理系统；2)高效的内存间数据共享和通信管理，保证混合NVM内存池中数据的一致性。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Today, enterprises are increasingly looking inward at their huge stores of under-processed or throwaway data, treating them as resources to be mined. As such, an emergent field called memory-intensive computing has ignited interest among industry and academia, largely driven by various emerging non-volatile memory technologies (NVMs). Machine Learning (ML) applications are being targeted by memory-intensive computing, leveraging unique properties of ML applications to improve their distributed performance by orders of magnitude. It is therefore highly desirable for ML applications executing in such memory intensive computing environments to be efficient, flexible and scalable. ML applications crunch a lot of data from disk drives, increasing latency due to disk access delays. This project explores and designs new techniques that let ML applications fully exploit the benefits of persistence for intermediate data in NVMs, which significantly reduces disk I/Os and hence data processing times. This project also involves curriculum development, and provides more avenues to bring women, minority, and underrepresented students into research and graduate programs.This project focuses on an open challenge for memory intensive computing systems how to offer ML applications with high efficiency, low cost and more flexibility, especially under the heterogeneous environment. This project proposes a hybrid NVM based computing architecture with effective data sharing and communication strategy to optimize file management, resource allocation and data communication for ML applications. This research centers on two key designs: 1) a new file and data management system based on the hybrid NVM pool consisting of Byte and Block addressable devices; 2) an efficient data sharing and communication management among memories to guarantee data consistency in the hybrid NVM memory pool.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.