arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1403.5337math.NAcs.NA
收藏

一种快速块低秩密集求解器及其在有限元矩阵中的应用

A Fast Block Low-Rank Dense Solver with Applications to Finite-Element Matrices

Stanford University(斯坦福大学);New York University(纽约大学)

Amirhossein Aminfar, Sivaram Ambikasaran, Eric Darve

AI总结:

本文提出了一种快速求解三维椭圆PDEs多前端稀疏消除过程中出现的密集 frontal 矩阵的求解器,通过高效表示为分层非对角低秩(HODLR)矩阵,采用边界距离低秩近似方法提升速度,适用于无结构网格,并在GMRES迭代中作为预条件子快速达到机器精度。

摘要:

本文提出了一种快速求解密集"frontal"矩阵的求解器,这些矩阵来源于三维椭圆PDEs的多前端稀疏消除过程。该求解器基于这些矩阵可以高效表示为分层非对角低秩(HODLR)矩阵的事实。为了构造非对角块的低秩近似,我们提出了一种新的伪骨架方案,即边界距离低秩近似,根据稀疏矩阵图中顶点的位置选择行和列。我们将这种新的低秩近似方法与自适应交叉近似(ACA)算法进行比较,证明其在无结构网格上具有更好的加速效果。使用HODLR直接求解器作为GMRES迭代方案的预条件子(低容差),可以比传统LU求解器更快达到机器精度。提供了针对三维有限元问题的frontal矩阵的数值基准,这些问题对应广泛的应用领域。

英文摘要:

This article presents a fast solver for the dense "frontal" matrices that arise from the multifrontal sparse elimination process of 3D elliptic PDEs. The solver relies on the fact that these matrices can be efficiently represented as a hierarchically off-diagonal low-rank (HODLR) matrix. To construct the low-rank approximation of the off-diagonal blocks, we propose a new pseudo-skeleton scheme, the boundary distance low-rank approximation, that picks rows and columns based on the location of their corresponding vertices in the sparse matrix graph. We compare this new low-rank approximation method to the adaptive cross approximation (ACA) algorithm and show that it achieves betters speedup specially for unstructured meshes. Using the HODLR direct solver as a preconditioner (with a low tolerance) to the GMRES iterative scheme, we can reach machine accuracy much faster than a conventional LU solver. Numerical benchmarks are provided for frontal matrices arising from 3D finite element problems corresponding to a wide range of applications.

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。

↑