MATEC Web Conf.
Volume 232, 20182018 2nd International Conference on Electronic Information Technology and Computer Engineering (EITCE 2018)
|Number of page(s)||6|
|Section||Circuit Simulation, Electric Modules and Displacement Sensor|
|Published online||19 November 2018|
A method for estimating existence of pairwise STM transaction conflicts
University of Novi Sad, Faculty of Technical Sciences, Trg Dositeja Obradovica 6, 21000 Novi Sad, Serbia
a Corresponding author: firstname.lastname@example.org
STM transaction schedulers were introduced to improve system performance. However, designing online transaction scheduling algorithms is challenging because at the same time they should: (i) introduce minimal scheduling overhead, (ii) minimize the resulting makespan, and (iii) minimize contention in the resulting schedule. In our previous work we developed the online transaction scheduler architecture and the four scheduling algorithms, named RR, ETLB, AC, and AAC (listed in increasing order of their quality), for scheduling transactions on the Python STM. Both AC and AAC use Bernstein conditions to check for pairwise data races between transactions, at the cost of time complexity that is proportional to the product of the sizes of transaction’s read and write sets, which may be significant. In this paper we propose a method for estimating existence of pairwise transaction conflicts whose time complexity is Θ(1). We validate this method by analysing the resulting transaction schedules for the three benchmark workloads, named RDW, CFW, and WDW. The result of this analysis is positive and encouraging – AAC using the new method produces the same result as when using Bernstein conditions. The limitation of the new method is that it may have false reports, both false negatives and false positives.
© The Authors, published by EDP Sciences, 2018
This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Current usage metrics show cumulative count of Article Views (full-text article views including HTML views, PDF and ePub downloads, according to the available data) and Abstracts Views on Vision4Press platform.
Data correspond to usage on the plateform after 2015. The current usage metrics is available 48-96 hours after online publication and is updated daily on week days.
Initial download of the metrics may take a while.