Analytical estimation of the scalability of iterative numerical algorithms on distributed memory multiprocessors

DOI10.1134/S1995080218040121zbMATH Open1398.65378arXiv1710.10490WikidataQ129778248 ScholiaQ129778248MaRDI QIDQ1662783FDOQ1662783

Authors: Y. Aharonov 

Publication date: 20 August 2018

Published in: Lobachevskii Journal of Mathematics (Search for Journal in Brave)

Abstract: This article presents a new high-level parallel computational model named BSF - Bulk Synchronous Farm. The BSF model extends the BSP model to deal with the compute-intensive iterative numerical methods executed on distributed-memory multiprocessor systems. The BSF model is based on the master-worker paradigm and the SPMD programming model. The BSF model makes it possible to predict the upper scalability bound of a BSF-program with great accuracy. The BSF model also provides equations for estimating the speedup and parallel efficiency of a BSF-program.

Full work available at URL: https://arxiv.org/abs/1710.10490

Recommendations

• On the speedup of parallel iterative numerical methods
• The performance of parallel iterative solvers
• The bulk-synchronous parallel random access machine
• Direct bulk-synchronous parallel algorithms
• A practical and portable model of programming for iterative solvers on distributed memory machines

zbMATH Keywords

iterative algorithmsdistributed memoryparallel computation modelBSF modelbulk synchronous farmscalability bound

Mathematics Subject Classification ID

Parallel numerical computation (65Y05) Numerical algorithms for specific classes of architectures (65Y10)

Cites Work

• H-BSP
• H-BSP: Hierarchical BSP computation model
• Title not available (Why is that?)
• Title not available (Why is that?)
• $\log_{\rm n}{\rm P}$ and $\log_{3}{\rm P}$: Accurate Analytical Models of Point-to-Point Communication in Distributed Systems
• A bridging model for multi-core computing
• The master-slave paradigm in parallel computer and industrial settings
• A single-program-multiple-data computational model for EPEX/FORTRAN
• A survey of the iterative methods for the solution of linear systems by extrapolation, relaxation and other techniques
• Title not available (Why is that?)

Cited In (3)

• Performance of iterative methods for distributed memory machines
• Scalability analysis of parallel GMRES implementations
• Scalable parallel algorithm for solving non-stationary systems of linear inequalities

Uses Software

• H-BSP

This page was built for publication: Analytical estimation of the scalability of iterative numerical algorithms on distributed memory multiprocessors

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1662783)