Collaborative Research: Frameworks: Basic ALgebra LIbraries for Sustainable Technology with Interdisciplinary Collaboration (BALLISTIC)

协作研究：框架：跨学科协作可持续技术的基本代数库（BALLISTIC）

• 批准号: 2004850
• 负责人: Julien Langou
• 金额: $ 64.62万
• 依托单位: University of Colorado at Denver-Downtown Campus
• 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: 2020-06-01 至 2024-05-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2004850&HistoricalAwards=false
• 关键词: Collaborative; Research; Frameworks; Basic; ALgebra

Scientific software libraries have long provided a large and growing resource for high-quality, reusable software components upon which applications from science and engineering can be rapidly constructed — with improved robustness, portability, and sustainability. For this, a team of researchers from four collaborating organizations proposed to develop BALLISTIC (Basic ALgebra Libraries for Sustainable Technology with Interdisciplinary Collaboration). The BALLISTIC project, through the leading-edge research it channels into its software deliverables, will lead to the introduction of tools that will simplify the transition to the next generation of extreme-scale computer architectures. The main impact of the project will be to develop, push, and deploy software into the scientific community to make it competitive on a world-wide scale and to contribute to standardization efforts in the area. BALLISTIC has the potential to become the community standard for dense linear algebra and be adopted and/or supported by a large community of users, computing centers, and High-Performance Computing (HPC) vendors. Learning to use numerical libraries is a basic part of the education of a computational scientist or engineer in many fields and at many academic institutions. BALLISTIC will have a correspondingly large impact on the research and education community, government laboratories, and private industry and support national efforts to build a workforce capable of employing state of the art tools in pursuit of science and engineering discoveries.The goal of BALLISTIC is to create a layered package of software components that is capable of running at every level of the platform deployment pyramid and achieves three complementary objectives: (1) deliver seamless access to the most up-to-date algorithms, numerics, and performance via familiar Sca/LAPACK interfaces, wherever possible; (2) make advanced algorithms, numerics, and performance capabilities available through new interface extensions, wherever necessary; and (3) provide a well-engineered conduit through which new discoveries at the frontiers of research in these areas can be channeled as quickly as possible to all the applications from science and engineering communities that depend on high-performance linear algebra libraries.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.

长期以来，科学软件库为高质量、可重用的软件组件提供了大量且不断增长的资源，在这些组件上可以快速构建来自科学和工程的应用程序--具有更高的健壮性、可移植性和可持续性。为此，来自四个合作组织的一组研究人员提议开发弹道学(具有跨学科合作的可持续技术基础代数库)。弹道项目通过其软件交付成果中的前沿研究，将导致工具的引入，这些工具将简化向下一代极端规模计算机体系结构的过渡。该项目的主要影响将是将软件开发、推广和部署到科学界，使其在世界范围内具有竞争力，并为该领域的标准化工作做出贡献。弹道学有可能成为密集线性代数的社区标准，并被大量用户、计算中心和高性能计算(HPC)供应商采用和/或支持。在许多领域和许多学术机构，学习使用数字图书馆是计算科学家或工程师教育的基本部分。弹道学将对研究和教育界、政府实验室和私营行业产生相应的巨大影响，并支持国家努力建立一支能够使用最先进工具追求科学和工程发现的劳动力队伍。弹道学的目标是创建能够在平台部署金字塔的每一层运行的分层软件组件包，并实现三个互补目标：(1)尽可能通过熟悉的SCA/LAPACK接口无缝访问最新的算法、数值和性能；(2)在必要时通过新的接口扩展提供先进的算法、数值和性能能力；以及(3)提供一个精心设计的渠道，通过该渠道，这些领域研究前沿的新发现可以尽快传送到依赖高性能线性代数库的科学和工程社区的所有应用程序。该奖项反映了NSF的法定使命，并通过使用基金会的智力优势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

Low-synch Gram–Schmidt with delayed reorthogonalization for Krylov solvers

具有延迟重新正交化的 Krylov 解算器的低同步 Gram-Schmidt

• DOI: 10.1016/j.parco.2022.102940
• 发表时间: 2022
• 期刊: Parallel Computing
• 影响因子: 1.4
• 作者: Bielich, Daniel;Langou, Julien;Thomas, Stephen;Świrydowicz, Kasia;Yamazaki, Ichitaro;Boman, Erik G.
• 通讯作者: Boman, Erik G.

Proposed Consistent Exception Handling for the BLAS and LAPACK

• DOI: 10.1109/correctness56720.2022.00006
• 发表时间: 2022-07
• 期刊: 2022 IEEE/ACM Sixth International Workshop on Software Correctness for HPC Applications (Correctness)
• 作者: J. Demmel;J. Dongarra;Mark Gates;G. Henry;J. Langou;Xiaoye Li;P. Luszczek;W. Pereira;Jason Riedy;Cindy Rubio-Gonz'alez

Symmetric Block-Cyclic Distribution: Fewer Communications Leads to Faster Dense Cholesky Factorization

对称块循环分布：更少的通信导致更快的密集 Cholesky 分解

• DOI: 10.1109/sc41404.2022.00034
• 发表时间: 2022
• 作者: Beaumont, Olivier;Duchon, Philippe;Eyraud-Dubois, Lionel;Langou, Julien;Vérité, Mathieu
• 通讯作者: Vérité, Mathieu

I/O-Optimal Algorithms for Symmetric Linear Algebra Kernels

对称线性代数内核的 I/O 优化算法

• DOI: 10.1145/3490148.3538587
• 发表时间: 2022
• 作者: Beaumont, Olivier;Eyraud-Dubois, Lionel;Langou, Julien;Vérité, Mathieu
• 通讯作者: Vérité, Mathieu