Show simple item record

dc.contributor.authorKyriacou, Costasen
dc.contributor.authorEvripidou, Paraskevasen
dc.contributor.authorTrancoso, Pedroen
dc.creatorKyriacou, Costasen
dc.creatorEvripidou, Paraskevasen
dc.creatorTrancoso, Pedroen
dc.description.abstractWith Data Driven Multithreading a thread is scheduled for execution only if all of its inputs have been produced and placed in the processor's local memory. Scheduling based on data availability may be used to exploit short-term optimal cache management policies. Such policies include firing a thread for execution only if its code and data are already placed in the cache. Furthermore, blocks associated to threads scheduled for execution in the near future, are not replaced until the thread starts its execution. We call this short-term optimal cache management policy the CacheFlow policy. Simulation results, on a 32-node system with CacheFlow, for eight scientific applications, have shown a significant reduction in the cache miss ratio. This results in an average speedup improvement of 18% when the basic prefetch CacheFlow policy is used, compared to the baseline data driven multithreading policy. This paper also presents two techniques to further improve the performance of CacheFlow: conflict avoidance and thread reordering. The results have shown an average speedup improvement of 26% and 31% for these two techniques, respectively. © Springer-Verlag 2004.en
dc.sourceLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)en
dc.subjectCache Missen
dc.subjectMicroprocessor chipsen
dc.subjectData availabilityen
dc.subjectData-driven multithreadingen
dc.subjectLocal memoryen
dc.subjectBaseline dataen
dc.subjectCache management policiesen
dc.subjectConflict avoidanceen
dc.subjectScientific applicationsen
dc.titleCacheFlow: A short-term optimal cache management policy for data driven multithreadingen
dc.description.endingpage570 Σχολή Θετικών και Εφαρμοσμένων Επιστημών / Faculty of Pure and Applied SciencesΤμήμα Πληροφορικής / Department of Computer Science
dc.description.notes<p>Cited By :9</p>en
dc.source.abbreviationLect. Notes Comput. Sci.en
dc.contributor.orcidTrancoso, Pedro [0000-0002-2776-9253]
dc.contributor.orcidEvripidou, Paraskevas [0000-0002-2335-9505]

Files in this item


There are no files associated with this item.

This item appears in the following Collection(s)

Show simple item record