# Efficient Scalable Algorithms for Solving Dense Linear Systems with Hierarchically Semiseparable Structures

@article{Wang2013EfficientSA,
title={Efficient Scalable Algorithms for Solving Dense Linear Systems with Hierarchically Semiseparable Structures},
author={Shen Wang and Xiaoye S. Li and Jianlin Xia and Yingchong Situ and Maarten V. de Hoop},
journal={SIAM J. Sci. Comput.},
year={2013},
volume={35}
}
• Published 11 December 2013
• Computer Science
• SIAM J. Sci. Comput.
Hierarchically semiseparable (HSS) matrix techniques are emerging in constructing superfast direct solvers for both dense and sparse linear systems. Here, we develop a set of novel parallel algorithms for key HSS operations that are used for solving large linear systems. These are parallel rank-revealing QR factorization, HSS constructions with hierarchical compression, ULV HSS factorization, and HSS solutions. The HSS tree-based parallelism is fully exploited at the coarse level. The…
42 Citations

## Figures and Tables from this paper

A Parallel Geometric Multifrontal Solver Using Hierarchically Semiseparable Structure
• Computer Science
ACM Trans. Math. Softw.
• 2016
A structured parallel geometry-based multifrontal sparse solver using hierarchically semiseparable (HSS) representations and exploiting the inherent low-rank structures is presented, which can be shown to have asymptotically lower complexities in both operation counts and memory than the conventional factorization algorithms for certain partial differential equations.
A Scalable Hierarchical Semi-Separable Library for Heterogeneous Clusters
• Computer Science
2017 46th International Conference on Parallel Processing (ICPP)
• 2017
This work presents algorithms for building and using HSS representations that are parallelized using MPI and CUDA to leverage state-of-the-art heterogeneous clusters and demonstrates excellent (linear) scalability on up to 128 GPUs on 128 nodes.
Parallel Randomized and Matrix-Free Direct Solvers for Large Structured Dense Linear Systems
• Computer Science
SIAM J. Sci. Comput.
• 2016
We design efficient and distributed-memory parallel randomized direct solvers for large structured dense linear systems, including a fully matrix-free version based on matrix-vector multiplications
Efficient Scalable Parallel Higher Order Direct MoM-SIE Method With Hierarchically Semiseparable Structures for 3-D Scattering
• Computer Science
IEEE Transactions on Antennas and Propagation
• 2017
The proposed novel fast scalable parallel algorithm enables dramatically faster monostatic scattering computations than iterative solvers and reduced number of unknowns when compared to low-order discretizations.
A Distributed-Memory Package for Dense Hierarchically Semi-Separable Matrix Computations Using Randomization
• Computer Science
ACM Trans. Math. Softw.
• 2016
A distributed-memory library for computations with dense structured matrices using Hierarchically Semi-Separable (HSS) representations and the compression algorithm that computes the HSS form of an input dense matrix relies on randomized sampling with a novel adaptive sampling mechanism.
Fast and scalable solvers for the Helmholtz equation
A new family of fast and scalable algorithms to solve the 2D high-frequency Helmholtz equation in heterogeneous medium based on a layered domain decomposition and a coupling between subdomains using the Green’s representation formula, which reduces the problem to a boundary integral system at the interfaces between subDomains.
A Roofline Performance Analysis of an Algebraic Multigrid PDE Solver
• Computer Science
• 2014
A performance analysis of a novel element-based algebraic multigrid (AMGe) method combined with a robust coarse-grid solution technique based on HSS lowrank sparse factorization found that keeping a small value of spectral tolerance is most critical to achieve the best AMG solver performance.
A Distributed-Memory Randomized Structured Multifrontal Method for Sparse Direct Solutions
• Computer Science
SIAM J. Sci. Comput.
• 2017
A distributed-memory randomized structured multifrontal solver for large sparse matrices that performs the major operations in terms of skinny matrices and fully structured forms significantly enhances the efficiency and scalability.
INV-ASKIT: A Parallel Fast Direct Solver for Kernel Matrices
• Computer Science
2016 IEEE International Parallel and Distributed Processing Symposium (IPDPS)
• 2016
A parallel algorithm for computing the approximate factorization of an N-by-N kernel matrix, which improves the state of the art by several orders of magnitude.

## References

SHOWING 1-10 OF 38 REFERENCES
Superfast Multifrontal Method for Large Structured Linear Systems of Equations
• Computer Science
SIAM J. Matrix Anal. Appl.
• 2009
A fast direct solver for large discretized linear systems using the supernodal multifrontal method together with low-rank approximations, especially suitable for large sparse problems and also has natural adaptability to parallel computations and great potential to provide effective preconditioners.
Fast algorithms for hierarchically semiseparable matrices
• Computer Science, Mathematics
Numer. Linear Algebra Appl.
• 2010
This paper generalizes the hierarchically semiseparable (HSS) matrix representations and proposes some fast algorithms for HSS matrices that are useful in developing fast‐structured numerical methods for large discretized PDEs, integral equations, eigenvalue problems, etc.
Efficient Structured Multifrontal Factorization for General Large Sparse Matrices
• J. Xia
• Computer Science
SIAM J. Sci. Comput.
• 2013
This work presents a framework of structured direct factorizations for general sparse matrices, including discretized PDEs on general meshes, based on the multifrontal method and hierarchically semiseparable (HSS) matrices.
On 3D modeling of seismic wave propagation via a structured parallel multifrontal direct Helmholtz solver
• Computer Science
• 2011
We consider the modeling of (polarized) seismic wave propagation on a rectangular domain via the discretization and solution of the inhomogeneous Helmholtz equation in 3D, by exploiting a parallel
Superfast and Stable Structured Solvers for Toeplitz Least Squares via Randomized Sampling
• Mathematics, Computer Science
SIAM J. Matrix Anal. Appl.
• 2014
This work generalizes standard hierarchically semiseparable (HSS) matrix representations to rectangular ones, and constructs a rectangular HSS approximation to $\mathcal{C}$ in nearly linear complexity with randomized sampling and fast multiplications of $\ mathcal{ C}$ with vectors.
Sweeping preconditioner for the Helmholtz equation: Hierarchical matrix representation
• Computer Science, Mathematics
• 2010
The sweeping preconditioner is introduced, which is highly efficient for iterative solutions of the variable‐coefficient Helmholtz equation including very‐high‐frequency problems and is extended to the three‐dimensional case with some success.
Massively parallel structured multifrontal solver for time-harmonic elastic waves in 3-D anisotropic media
• Computer Science, Geology
• 2012
SUMMARY We present a massively parallel structured multifrontal solver for the equations describing time-harmonic elastic waves in 3-D anisotropic media. We use a multicomponent second-order
Massively parallel structured direct solver for equations describing time-harmonic qP-polarized waves in TTI media
• Geology
• 2012
ABSTRACTWe considered the discretization and approximate solutions of equations describing time-harmonic qP-polarized waves in 3D inhomogeneous anisotropic media. The anisotropy comprises general
A Fast ULV Decomposition Solver for Hierarchically Semiseparable Representations
• Mathematics, Computer Science
SIAM J. Matrix Anal. Appl.
• 2006
We consider an algebraic representation that is useful for matrices with off-diagonal blocks of low numerical rank. A fast and stable solver for linear systems of equations in which the coefficient