A Complexity-Based Hierarchy for Multiprocessor Synchronization
Open Access
- 25 July 2016
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 289-298
Abstract
For many years, Herlihy's elegant computability based Consensus Hierarchy has been our best explanation of the relative power of various types of multiprocessor synchronization objects when used in deterministic algorithms. However, key to this hierarchy is treating these instructions as distinct objects, an approach that is far from the real-world, where multiprocessor programs apply synchronization instructions to collections of arbitrary memory locations. We were surprised to realize that, when considering instructions applied to memory locations, the computability based hierarchy collapses. This leaves open the question of how to better captures the power of various synchronization instructions. In this paper, we provide an approach to answering this question. We present a hierarchy of synchronization instructions, classified by their space complexity in solving obstruction-free consensus. Our hierarchy provides a classification of combinations of known instructions that seems to fit with our intuition of how useful some are in practice, while questioning the effectiveness of others. We prove an essentially tight characterization of the power of buffered read and write instructions. Interestingly, we show a similar result for multi-location atomic assignments.Keywords
All Related Versions
Funding Information
- Natural Sciences and Engineering Research Council of Canada (RGPIN-2015-05080)
- Department of Energy (ER26116/DE-SC0008923)
- National Science Foundation (CCF-1217921,CCF-1301926,IIS-1447786)
This publication has 18 references indexed in Scilit:
- On the Optimal Space Complexity of Consensus for Anonymous Processes2015
- An $O(\sqrt n)$ Space Bound for Obstruction-Free Leader ElectionLecture Notes in Computer Science, 2013
- Obstruction-Free Algorithms Can Be Practically Wait-FreeLecture Notes in Computer Science, 2005
- What Can Be Implemented Anonymously?Lecture Notes in Computer Science, 2005
- Determining Consensus NumbersSIAM Journal on Computing, 2000
- All of Us Are Smarter than Any of Us: Nondeterministic Wait-Free Hierarchies Are Not RobustSIAM Journal on Computing, 2000
- On the space complexity of randomized synchronizationJournal of the ACM, 1998
- Atomic snapshots of shared memoryJournal of the ACM, 1993
- Wait-free synchronizationACM Transactions on Programming Languages and Systems, 1991
- Fast randomized consensus using shared memoryJournal of Algorithms, 1990