Optimal Parallel Solution of Sparse Triangular Systems
نویسندگان
چکیده
This paper describes a method for the parallel solution of tr/ans_lar sets of equations, appropriate when there are many rlght-hand sides. By preproceuing, the method can reduce the number of parsl/el steps required to solve Zz = b compared to parallel forward or backsolve. AppUcations are to iterative solvers with tzian_d&r preconditioners, to structural sns]ysds, or to power system= applications, where there may be many right-hand sides (not all available = p_o_. The inverse of £ is represented as a product of sparse trian_als_ factors. The problem considered in this paper is to find a factored representation of this inverse of L with the smsllest number of factors "This work was supported in part under NSF Contracts ECS-8822654 and ECS8907391. tThis work was supported by the NAS System= Division and/or DARPA via Cooperstire A_.ement NCC 2-387 between NASA and the University Space R_esrch Association (USR.A). Work was performed st the Research Institute for Advanced Computer Science (RIACS), NASA Ames Research Center, Moffett Field, CA 94035 and at CERFACS, 42 Ave. Gustave Coriolis, 31057 Toulouse Cedex, France. The assistance of CE_ACS and its staff during this author's visit in Au4just 1989 are gratefully acknowledged.
منابع مشابه
Optimal Dag Partitioning for Partially Inverting Triangular Systems
An approach for solving sparse triangular systems of equations on highly parallel computers employs a partitioned representation of the inverse of the triangular matrix so that the solution can be obtained by a series of matrix-vector multiplications. This approach requires a number of global communication steps that is proportional to the number of factors in the partitioning. The problem of n...
متن کاملIterative Sparse Triangular Solves for Preconditioning
Sparse triangular solvers are typically parallelized using levelscheduling techniques, but parallel efficiency is poor on high-throughput architectures like GPUs. We propose using an iterative approach for solving sparse triangular systems when an approximation is suitable. This approach will not work for all problems, but can be successful for sparse triangular matrices arising from incomplete...
متن کاملStructure-adaptive parallel solution of sparse triangular linear systems
Solving sparse triangular systems of linear equations is a performance bottleneck in many methods for solving more general sparse systems. Both for direct methods and for many iterative preconditioners, it is used to solve the system or improve an approximate solution, often across many iterations. Solving triangular systems is notoriously resistant to parallelism, however, and existing paralle...
متن کاملParallel Solution of Sparse Triangular Linear Systems in the Preconditioned Iterative Methods on the GPU
A novel algorithm for solving in parallel a sparse triangular linear system on a graphical processing unit is proposed. It implements the solution of the triangular system in two phases. First, the analysis phase builds a dependency graph based on the matrix sparsity pattern and groups the independent rows into levels. Second, the solve phase obtains the full solution by iterating sequentially ...
متن کاملParallel Solution of Sparse Linear Least Squares Problemson Distributed - Memory
This paper studies the solution of large-scale sparse linear least squares problems on distributed-memory multiprocessors. The method of corrected semi-normal equations is considered. New block-oriented parallel algorithms are developed for solving the related sparse triangular systems. The arithmetic and communication complexities of the new algorithms applied to regular grid problems are anal...
متن کاملA high performance two dimensional scalable parallel algorithm for solving sparse triangular systems
Solving a system of equations of the form Tx = y, where T is a sparse triangular matrix, is required after the factorization phase in the direct methods of solving systems of linear equations. A few parallel formulations have been proposed recently. The common belief in parallelizing this problem is that the parallel formulation utilizing a two dimensional distribution of T is unscalable. In th...
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
عنوان ژورنال:
- SIAM J. Scientific Computing
دوره 14 شماره
صفحات -
تاریخ انتشار 1993