Paper Info
Title
Hardware Transactional Memory Based on Abort Prediction and Adaptive Retry Policy for Multi-Core In-Memory Databases
Abstract
Since Intel has recently shifted Transactional Synchronization Extension (TSX) as its first mainstream Hardware Transactional Memory (HTM), HTM has greatly changed the parallel programming paradigm for transaction processing, As a result, a number of studies on HTM have been conducted actively. However, the existing studies consider only the prediction of a conflict between two transactions and provide a static HTM configuration for all workloads. To solve the problems, we propose an efficient hardware transactional memory scheme based on both abort prediction and adaptive retry policy for multi-core in-memory databases. First, the proposed scheme can predict not only conflicts between transactions running concurrently, but also the capacity and other aborts of transactions by collecting the information of previously executed transactions. Second, the proposed scheme can provide a near-optimal HTM configuration according to the characteristic of a given workload by using an adaptive retry policy based on machine learning algorithms. Finally, through our experimental performance analysis using STAMP, the proposed scheme shows about 30~40% better performance than the existing HTM-based schemes.
Year
DOI
Venue
2018
10.1109/BigComp.2018.00061
2018 IEEE International Conference on Big Data and Smart Computing (BigComp)
Keywords
Field
DocType
Hardware Transactional Memory(HTM),abort prediction,retry policy,multi-core in memory database
Abort,Resource management,Transaction processing,Synchronization,Computer science,Instruction set,Transactional memory,Memory management,Multi-core processor,Database
Conference
ISSN
ISBN
Citations 
2375-933X
978-1-5386-3650-3
0
PageRank 
References 
Authors
0.34
0
5
Name
Order
Citations
PageRank
Hyeong-Jin Kim100.34
Mun-Hwan Kang200.34
Yeon-Woo Chang300.34
Min Yoon43410.38
Jae-Woo Chang540199.85