Results 1 -
9 of
9
A Performance Analysis of Transposition-Table-Driven Scheduling in Distributed Search
- IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
, 2002
"... 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 proce ..."
Abstract
-
Cited by 13 (8 self)
- Add to MetaCart
(Show Context)
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.
Distributed game-tree search using transposition table driven work scheduling
- in International Conference on Parallel Processing (ICPP), 2002
"... Abstract The fffi algorithm for two-player game-tree search hasa notorious reputation as being a challenging algorithm for achieving reasonable parallel performance. MTD(f),a new fffi variant, has become the sequential algorithmof choice for practitioners. Unfortunately, MTD(f) inherits most of the ..."
Abstract
-
Cited by 5 (1 self)
- Add to MetaCart
(Show Context)
Abstract The fffi algorithm for two-player game-tree search hasa notorious reputation as being a challenging algorithm for achieving reasonable parallel performance. MTD(f),a new fffi variant, has become the sequential algorithmof choice for practitioners. Unfortunately, MTD(f) inherits most of the parallel obstacles of fffi, as well as creat-ing new performance hurdles. Transposition-table-driven scheduling (TDS) is a new parallel search algorithm thathas proven to be effective in the single-agent (one-player) domain. This paper presents TDSAB, the first time TDS par-allelism has been applied to two-player search (the MTD(f) algorithm). Results show that TDSAB gives comparablespeedups to that achieved by conventional parallel fffi al-gorithms. However, since this is a parallelization of a superior sequential algorithm, the results in fact are better. Thispaper shows that the TDS idea can be extended to more challenging search domains.Keywords: fffi search, transposition-table-driven schedul-ing, single-agent search, transposition table.
Wide-Area Transposition-Driven Scheduling
- In IEEE International Symposium on High Performance Distributed Computing
, 2001
"... Distributed search of state spaces containing cycles is a challenging task and has been studied for years. Traditional parallel search algorithms either ignore the cyclic nature of the state space and waste much time in duplicated search effort, or rely on heavy communication to reduce duplicate wor ..."
Abstract
-
Cited by 4 (2 self)
- Add to MetaCart
(Show Context)
Distributed search of state spaces containing cycles is a challenging task and has been studied for years. Traditional parallel search algorithms either ignore the cyclic nature of the state space and waste much time in duplicated search effort, or rely on heavy communication to reduce duplicate work, resulting in a large communication overhead. Both methods perform poorly, even when using a fast, local interconnect. A recently developed task-distribution scheme, Transposition-Driven Scheduling (TDS), performs much better, since it communicates asynchronously and efficiently suppresses duplicate search effort. TDS, however, requires bandwidths of megabytes per second per processor. In this paper, we investigate how cyclic state spaces can be searched efficiently on a meta-computing system containing multiple clusters, connected by high-latency, lowbandwidth wide-area links. This is quite a challenge, because the wide-area links provide neither the bandwidth required for TDS, nor the latency required for traditional distributed search algorithms. We propose a scheme that strongly reduces communication between clusters, at the expense of possibly duplicate search effort. Performance measurements for several applications show that the new scheme outperforms traditional schemes by a wide margin. Keywords: meta computing, distributed search, distributed supercomputing, Transposition-Driven Scheduling 1
Transposition-Driven Scheduling in Parallel Two-Player State-Space Search
, 2005
"... This thesis describes the design and implementation of a parallel two-player statespace searcher called DARKSIGHT. The program uses principal variation search as its main search algorithm with numerous additional techniques and heuristics. The young-brothers-wait concept is used as the splitting str ..."
Abstract
-
Cited by 1 (0 self)
- Add to MetaCart
This thesis describes the design and implementation of a parallel two-player statespace searcher called DARKSIGHT. The program uses principal variation search as its main search algorithm with numerous additional techniques and heuristics. The young-brothers-wait concept is used as the splitting strategy. Transpositiondriven scheduling is used as its load-balancing scheme, which achieved near-linear speedups in parallel one-player state-space search. Our experiments show that transposition-driven scheduling is used successful with performance results that are comparable to those achieved by other parallel two-player state-space searchers. Additionally, we present new experimental results on the deep search behaviour of two-player state-space searchers, and results that indicate that the commonly used constant-sized aspiration windows are not optimal.
Transposition Table Driven Scheduling for Two-Player Games
, 2002
"... Game-tree search is an important research topic in Artificial Intelligence. Because it is computationally intensive, researchers have turned their at- tention to parallel game-tree search algorithms in order to improve running time. However, achieving high parallel performance remains a difficult ta ..."
Abstract
-
Cited by 1 (1 self)
- Add to MetaCart
Game-tree search is an important research topic in Artificial Intelligence. Because it is computationally intensive, researchers have turned their at- tention to parallel game-tree search algorithms in order to improve running time. However, achieving high parallel performance remains a difficult task on distributed-memory systems. Many high-performance systems for two- player games use the c/ search algorithm, enhanced with transposition tables. Transposition tables store useful information about game-trees from previous searches. When parallelizing the c algorithm, a major problem is sharing transposition table information efficiently among the processors. A processor must communicate in order to look up table entries on other processors, which hurts the performance of the parallel search.
The Multigame Reference Manual
, 2000
"... This document describes Multigame, a language for expressing the rules of one-person and two-person board games, such as chess, checkers, and the 15-puzzle. The programmer writes the rules of a game in a Multigame program. From this description, the Multigame compiler generates a (parallel) program ..."
Abstract
-
Cited by 1 (1 self)
- Add to MetaCart
(Show Context)
This document describes Multigame, a language for expressing the rules of one-person and two-person board games, such as chess, checkers, and the 15-puzzle. The programmer writes the rules of a game in a Multigame program. From this description, the Multigame compiler generates a (parallel) program that can play the game. The document gives both an informal introduction to the language and a formal definition of the grammar. The document also describes how a Multigame playing program communicates moves in a nearly game-independent way. To improve the quality of the game-playing program, the programmer can provide an evaluation function with game-dependent heuristics. The evaluation function should be written in C; we describe the interface in this document as well. 1 Introduction Multigame is an application-domain-oriented, high-level language for describing one-person and twoperson board games. A Multigame program describes the legal moves of a game in a formal way. The Multigame fro...
Applying the Experience of Building a High-Performance Search Engine for One Domain to Another
, 2002
"... This paper describes the high-performance alpha-beta-based search engine used in CHINOOK, the World Man-Machine Checkers Champion. Previous experience in designing a chess program was important in the initial design of the program. As it evolved, however, numerous application-specific modifications ..."
Abstract
- Add to MetaCart
This paper describes the high-performance alpha-beta-based search engine used in CHINOOK, the World Man-Machine Checkers Champion. Previous experience in designing a chess program was important in the initial design of the program. As it evolved, however, numerous application-specific modifications had to be made for the algorithms to excel at searching checkers trees. This paper describes the experience of transferring the technology used to develop a chess program to the creation of a high-performance checkers program.
by
, 2007
"... This dissertation, written by David M. Kaiser, and entitled The Structure of Games, having been approved in respect to style and intellectual content, is referred to you for judgment. We have read this dissertation and recommend that it be approved. ..."
Abstract
- Add to MetaCart
(Show Context)
This dissertation, written by David M. Kaiser, and entitled The Structure of Games, having been approved in respect to style and intellectual content, is referred to you for judgment. We have read this dissertation and recommend that it be approved.
