Load Balancing Problem for Parallel Computers with Distributed Memory
نویسندگان
چکیده
This paper deals with load balancing of parallel algorithms for distributedmemory computers. The parallel versions of BLAS subroutines for matrix-vector product and LU factorization are considered. Two task partitioning algorithms are investigated and speed-ups are calculated. The cases of homogeneous and heterogeneous collections of computers/processors are studied, and special partitioning algorithms for heterogeneous workstation clusters are presented.
منابع مشابه
Static versus dynamic heterogeneous parallel schemes to solve the symmetric tridiagonal eigenvalue problem
Computation of the eigenvalues of a symmetric tridiagonal matrix is a problem of great relevance. Many linear algebra libraries provide subroutines for solving it. But none of them is oriented to be executed in heterogeneous distributed memory multicomputers. In this work we focus on this kind of platforms. Two different load balancing schemes are presented and implemented. The experimental res...
متن کاملClassification of Load Balancing Conditions for parallel and distributed systems
Although intensive work has been done in the area of load balancing, the measure of success of load balancing is the net execution time achieved by applying the load balancing algorithms. This paper deals with the problem of load balancing conditions of parallel and distributed applications. Parallel and distributed computers have multiple-CPU architecture, and in parallel system they have shar...
متن کاملA Fully Parallel Condensation Method for Generalized Eigenvalue Problems on Distributed Memory Computers
For large-scale eigenvalue problems the authors introduced in 16] a coarse grained parallel algorithm for distributed memory computers based on sub-structuring and improved static condensation. It was observed that the load balancing may become unsatisfactory if the master problem and the problems corresponding to the substructures are of signiicantly diierent size. In this paper we improve the...
متن کاملMolecular Dynamics with Load Balancing on Distributed - Memory Mimd Computers
We report two aspects of a computational molecular dynamics study of large-scale problems on a distributed-memory MIMD parallel computer: (1) efficiency and scalability results on Intel Paragon parallel computers with up to 512 nodes and (2) a new method for dynamic load balancing.
متن کاملRunning Finite-difference Schemes for 3d Diffusion Problems on Parallel Computers with Distributed Memory
In this papper we consider the problem of solving 3D diffusion problems on distributed memory computers. We present a parallel algorithm that is suitable for the number of processors less or equal 8. The pipelining method is used to enlarge the number of processors till 64. The computational grid decomposition method is proposed for heterogenous clusters of workstations which preserves the load...
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 2014