RT Conference Proceedings
T1 Improving Transactional Memory Performance for Irregular Applications
A1 Pedrero-Luque, Manuel
A1 Gutiérrez-Carrasco, Eladio Damián
A1 Romero-Montiel, Sergio
A1 Plata-González, Óscar Guillermo
K1 Arquitectura de ordenadores
AB Transactional memory (TM) offers optimistic concurrency support in modern multicore archi-tectures, helping the programmers to extract parallelism in irregular applications when datadependence information is not available before runtime. In fact, recent research focus on ex-ploiting thread-level parallelism using TM approaches. However, the proposed techniques areof general use, valid for any type of application.This work presents ReduxSTM, a software TM system specially designed to extract maxi-mum parallelism from irregular applications. Commit management and conflict detection aretailored to take advantage of both, sequential transaction ordering to assure correct results,and privatization of reduction patterns, a very frequent memory access pattern in irregularapplications. Both techniques are used to avoid unnecessary transaction aborts.A function in 300.twolf package from SPEC CPU2000 was taken as a motivating irregularprogram. This code was parallelized using ReduxSTM and an ordered version of TinySTM,a state-of-the-art TM system. Experimental evaluation shows that ReduxTM exploits moreparallelism from the sequential program and obtains better performance than the other system.
YR 2015
FD 2015-06-11
LK http://hdl.handle.net/10630/9889
UL http://hdl.handle.net/10630/9889
LA eng
NO Postprint de autor publicado posteriormente con este DOI:http://dx.doi.org/10.1016/j.procs.2015.05.398
NO Universidad de Málaga. Campus de Excelencia Internacional Andalucía Tech.
DS RIUMA. Repositorio Institucional de la Universidad de Málaga
RD 22 ene 2026