A Performance Analysis of Transposition-Table-Driven Work Scheduling in Distributed Search

J.W. Romein, H.E. Bal, J. Schaeffer, A. Plaat

Research output: Contribution to JournalArticleAcademicpeer-review

Abstract

This paper discusses a new work-scheduling algorithm for parallel search of single-agent state spaces, called Transposition-Table-Driven Work Scheduling, that places the transposition table at the heart of the parallel work scheduling. The scheme results in less synchronization overhead, less processor idle time, and less redundant search effort. Measurements on a 128-processor parallel machine show that the scheme achieves close-to-linear speedups; for large problems the speedups are even superlinear due to better memory usage. On the same machine, the algorithm is 1.6 to 12.9 times faster than traditional work-stealing-based schemes.
Original languageEnglish
Pages (from-to)447-459
JournalIEEE Transactions on Parallel and Distributed Systems
Volume13
Issue number5
DOIs
Publication statusPublished - 2002

Bibliographical note

tds:tpds

Fingerprint Dive into the research topics of 'A Performance Analysis of Transposition-Table-Driven Work Scheduling in Distributed Search'. Together they form a unique fingerprint.

Cite this