arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1608.00044cs.MS
收藏

基于异步任务的扇双稀疏Cholesky求解器

An Asynchronous Task-based Fan-Both Sparse Cholesky Solver

Mathias Jacquelin, Yili Zheng, Esmond Ng, Katherine Yelick

摘要:

线性方程组是许多科学和工程应用的核心。这些线性系统中的许多是稀疏的;即,系数矩阵中的大多数元素为零。有时需要基于矩阵分解的直接方法来确保精确的解。例如,移位-逆Lanczos算法需要精确求解稀疏线性系统的内部特征值。稀疏矩阵分解的性能和资源使用对于解决时间和在给定平台上可解决的最大问题大小至关重要。在许多应用中,系数矩阵是对称的,并且利用对称性将减少因式分解所需的工作量和存储成本。在大规模分布式存储平台上进行因式分解时,通信开销对算法性能的影响至关重要。与此同时,网络拓扑结构变得越来越复杂,因此现代平台表现出很高的性能可变性。这使得计算调度成为一项复杂且性能关键的任务。在本文中,我们调查使用的异步任务范式,单边通信和动态调度在实现稀疏Cholesky分解(symPACK)的大规模分布式内存平台。我们的求解器symPACK依赖于UPC++库提供的高效灵活的通信原语。性能评估显示了良好的可扩展性,symPACK的性能优于最先进的并行分布式内存分解包,验证了我们的方法在实际情况下。

英文摘要:

Systems of linear equations arise at the heart of many scientific and engineering applications. Many of these linear systems are sparse; i.e., most of the elements in the coefficient matrix are zero. Direct methods based on matrix factorizations are sometimes needed to ensure accurate solutions. For example, accurate solution of sparse linear systems is needed in shift-invert Lanczos to compute interior eigenvalues. The performance and resource usage of sparse matrix factorizations are critical to time-to-solution and maximum problem size solvable on a given platform. In many applications, the coefficient matrices are symmetric, and exploiting symmetry will reduce both the amount of work and storage cost required for factorization. When the factorization is performed on large-scale distributed memory platforms, communication cost is critical to the performance of the algorithm. At the same time, network topologies have become increasingly complex, so that modern platforms exhibit a high level of performance variability. This makes scheduling of computations an intricate and performance-critical task. In this paper, we investigate the use of an asynchronous task paradigm, one-sided communication and dynamic scheduling in implementing sparse Cholesky factorization (symPACK) on large-scale distributed memory platforms. Our solver symPACK relies on efficient and flexible communication primitives provided by the UPC++ library. Performance evaluation shows good scalability and that symPACK outperforms state-of-the-art parallel distributed memory factorization packages, validating our approach on practical cases.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑