Collaborative Research: Elements: ProDM: Developing A Unified Progressive Data Management Library for Exascale Computational Science

协作研究：要素：ProDM：为百亿亿次计算科学开发统一的渐进式数据管理库

• 批准号: 2311758
• 负责人: Xubin He
• 金额: $ 18万
• 依托单位: Temple University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2023
• 资助国家: 美国
• 起止时间: 2023-08-01 至 2026-07-31
• 项目状态: 未结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2311758&HistoricalAwards=false
• 关键词: Collaborative; Research; Elements; ProDM; Developing

Effective management of scientific data produced by extreme-scale simulations and instruments is crucial for advancing scientific discoveries. Due to the scale of data and the diverse requirements of scientific analytics, there is a growing need to manage data in a progressive manner, such that users can stream as much data as they need to carry out their data analytics with reduced data movement and computation. However, little effort has been put into creating robust and scalable cyberinfrastructure services that link the recent algorithmic innovations in progressive methods with scientific data analytics, leaving these capabilities inaccessible to scientists. This project aims to develop a sustainable framework ProDM that supports the progressive management of scientific data to facilitate its use in scientific applications. The success of this project will enable new scientific research and novel findings by providing a new way to manage and analyze data. Furthermore, outcomes of this project will be delivered as publicly available software to enhance research cyberinfrastructure, promote education and teaching, and broaden participation in computing. ProDM is centered upon the unification of viable progressive representations and tailored development for in-situ and post-hoc analytic routines. In particular, it involves three key activities. First, a data engine will be built to unify state-of-the-art progressive representations, and provide portable hardware support for accelerators as well as interoperative software interfaces to other data management and analytic libraries. Second, an in-situ engine will be developed to facilitate the use of progressive representations for in-situ data analytics, which include a redesign of in-situ semantics and adjustment of runtime dynamics. Third, a post-hoc engine will be developed to efficiently access progressive data and improve the performance of data retrieval for post-hoc data analytics. ProDM will be deployed on campus-wide computing infrastructures and leadership systems for integration and evaluation with real-world scientific applications from climate, fusion, molecular dynamics, and beyond.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.

有效管理由超大规模模拟和仪器产生的科学数据对于推进科学发现至关重要。由于数据规模和科学分析的多样化要求，越来越需要以渐进的方式管理数据，以便用户可以流式传输执行数据分析所需的尽可能多的数据，同时减少数据移动和计算。然而，在创建强大且可扩展的网络基础设施服务方面却投入很少，这些服务将渐进方法中的最新算法创新与科学数据分析联系起来，导致科学家无法获得这些功能。该项目旨在开发一个可持续的框架 ProDM，支持科学数据的渐进式管理，以促进其在科学应用中的使用。该项目的成功将提供一种管理和分析数据的新方法，从而实现新的科学研究和新发现。此外，该项目的成果将作为公开软件提供，以增强研究网络基础设施、促进教育和教学并扩大对计算的参与。 ProDM 的重点是可行的渐进表示的统一以及针对现场和事后分析例程的定制开发。具体而言，它涉及三项关键活动。首先，将构建一个数据引擎来统一最先进的渐进表示，并为加速器提供便携式硬件支持以及与其他数据管理和分析库的互操作软件接口。其次，将开发一个原位引擎，以促进使用渐进式表示进行原位数据分析，其中包括重新设计原位语义和调整运行时动态。第三，将开发事后引擎来有效访问渐进数据并提高事后数据分析的数据检索性能。 ProDM 将部署在校园范围内的计算基础设施和领导系统上，用于与气候、聚变、分子动力学等领域的现实科学应用进行集成和评估。该奖项反映了 NSF 的法定使命，并被认为值得通过使用评估来支持基金会的智力价值和更广泛的影响审查标准。

项目成果

Improving Progressive Retrieval for HPC Scientific Data using Deep Neural Network

使用深度神经网络改进 HPC 科学数据的渐进检索

• DOI: 10.1109/icde55515.2023.00209
• 发表时间: 2023-04
• 期刊: IEEE
• 作者: Wang, Jinzhen;Liang, Xin;Whitney, Ben;Chen, Jieyang;Gong, Qian;He, Xubin;Wan, Lipeng;Klasky, Scott;Podhorszki, Norbert;Liu, Qing
• 通讯作者: Liu, Qing