Minimizing the memory usage with out-of-core multi-frontal direct solver
نویسندگان
چکیده
In this paper we present the out-of-core multi-frontal direct solver algorithm for multi-physics problems. The solver is based on the binary elimination tree, with number of leaves equal to the number of finite elements. The solver browses the elimination tree in the post-order, eliminates fully assembled degrees of freedom and merges resulting Schur complement matrices. The solver is intended to minimize the memory usage by dumping out all Schur complements and processed local systems to the disc. The solver algorithm is executed on challenging computational problem involving the linear elasticity coupled with linear accoustics. The history of execution of the solver algorithm is analyzed by measuring the memory usage of the solver as well as the sparsities of the nodal matrices. The memory usage of the solver is compared with the MUMPS solver.
منابع مشابه
Reducing the I/O Volume in an Out-of-Core Sparse Multifrontal Solver
High performance sparse direct solvers are often a method of choice in various simulation problems. However, they require a large amount of memory compared to iterative methods. In this context, out-of-core solvers must be employed, where disks are used when the storage requirements are too large with respect to the physical memory available. In this paper, we study how to minimize the I/O requ...
متن کاملA Preliminary Out-of-Core Extension of a Parallel Multifrontal Solver
The memory usage of sparse direct solvers can be the bottleneck to solve large-scale problems. This paper describes a first implementation of an out-of-core extension to a parallel multifrontal solver (MUMPS). We show that larger problems can be solved on limited-memory machines with reasonable performance, and we illustrate the behaviour of our parallel out-of-core factorization. Then we use s...
متن کاملParallel multifrontal method with out-of-core techniques
The memory usage of sparse direct solvers can be the bottleneck to solve large-scale problems. We describe the ongoing work on the implementation of an out-of-core extension to a parallel multifrontal solver (MUMPS). We show that larger problems can be solved on limited-memory machines with reasonable performance, and we illustrate the behaviour of both the parallel out-of-core factorization an...
متن کاملOptimal memory minimization algorithms for the multifrontal method
We are interested in the active and total memory usage of the mul-tifrontal method. Starting from the algorithms proposed by Liu, we suggest a new scheme together with a tree traversal that give an optimal peak of active memory. Significant gains are obtained compared to Liu's approach. We also study the problem of minimizing the total memory and compare various new schemes. A number of experim...
متن کاملComputational complexity and memory usage for multi-frontal direct solvers in structured mesh finite elements
The multi-frontal direct solver is the state-of-the-art algorithm for the direct solution of sparse linear systems. This paper provides computational complexity and memory usage estimates for the application of the multifrontal direct solver algorithm on linear systems resulting from B-spline-based isogeometric finite elements, where the mesh is a structured grid. Specifically we provide the es...
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:
دوره شماره
صفحات -
تاریخ انتشار 2011