Lifting the barriers --- reducing latencies with transparent transactional memory

Authors:
Annette Bieniusa;Thomas Fuhrmann
Affiliations:
Department for Computer Science, Universität Freiburg, Freiburg, Germany;Department of Informatics, Technische Universität München, Garching, Germany
Venue:
ICDCN'12 Proceedings of the 13th international conference on Distributed Computing and Networking
Year:
2012

Citing 10
Cited 1

Java(TM) Language Specification, The (3rd Edition) (Java (Addison-Wesley))

Java(TM) Language Specification, The (3rd Edition) (Java (Addison-Wesley))
Eliminating synchronization-related atomic operations with biased locking and bulk rebiasing

Proceedings of the 21st annual ACM SIGPLAN conference on Object-oriented programming systems, languages, and applications
Speculative execution in a distributed file system

ACM Transactions on Computer Systems (TOCS)
A Uniform Transactional Execution Environment for Java

ECOOP '08 Proceedings of the 22nd European conference on Object-Oriented Programming
Stretching transactional memory

Proceedings of the 2009 ACM SIGPLAN conference on Programming language design and implementation
Transactional Memory, 2nd Edition

Transactional Memory, 2nd Edition
The 48-core SCC Processor: the Programmer's View

Proceedings of the 2010 ACM/IEEE International Conference for High Performance Computing, Networking, Storage and Analysis
Transactions as the foundation of a memory consistency model

DISC'10 Proceedings of the 24th international conference on Distributed computing
The architecture of the DecentVM: towards a decentralized virtual machine for many-core computing

Virtual Machines and Intermediate Languages
Transactional locking II

DISC'06 Proceedings of the 20th international conference on Distributed Computing

WTTM 2011: the third workshop on the theory of transactional memory

ACM SIGACT News

Quantified Score

Hi-index	0.00

Visualization

Abstract

Synchronization in distributed systems is expensive because, in general, threads must stall to obtain a lock or to operate on volatile data. Transactional memory, on the other hand, allows speculative execution so that it can hide the latencies that are inherent to distributed systems. In this paper, we discuss how transactional memory can carry over to code that uses Java's synchronization means i.e. monitors and volatiles. We show that we can guarantee correct execution according to the Java memory model (JMM) without having to stall at synchronization points. To this end, we use a multi-version software transactional memory system that executes JMM synchronization operations asynchronously. If any such execution has violated the JMM, the transaction rolls back. As a result, only blocking operations require immediate synchronization barriers.