Title :
Employing Transactional Memory and Helper Threads to Speedup Dijkstra´s Algorithm
Author :
Nikas, Konstantinos ; Anastopoulos, Nikos ; Goumas, Georgios ; Koziris, Nectarios
Author_Institution :
Comput. Syst. Lab., Nat. Tech. Univ. of Athens, Athens, Greece
Abstract :
In this paper we work on the parallelization of the inherently serial Dijkstra´s algorithm on modern multicore platforms. Dijkstra´s algorithm is a greedy algorithm that computes single source shortest paths for graphs with non-negative edges and is based on the iterative extraction of nodes from a priority queue. This property limits the explicit parallelism of the algorithm and any attempt to utilize the remaining parallelism results in significant slowdowns due to synchronization overheads. To deal with these problems, we employ the concept of helper threads (HT) to extract parallelism on a non-traditional fashion and transactional memory (TM) to efficiently orchestrate the concurrent threads´ accesses to shared data structures. Results demonstrate that the proposed implementation is able to achieve performance speedups (reaching up to 1.84 for 14 threads), indicating that the two paradigms could be efficiently combined.
Keywords :
greedy algorithms; microcomputers; Dijkstra algorithm; greedy algorithm; helper threads; iterative node extraction; multicore platforms; single source shortest paths; transactional memory; Concurrent computing; Data mining; Greedy algorithms; Iterative algorithms; Laboratories; Multicore processing; Parallel processing; Programming profession; Systems engineering and theory; Yarn; Graph Algorithms; Helper Threading; SSSP algorithms; Speculative Multithreading; Transactional Memory;
Conference_Titel :
Parallel Processing, 2009. ICPP '09. International Conference on
Conference_Location :
Vienna
Print_ISBN :
978-1-4244-4961-3
Electronic_ISBN :
0190-3918
DOI :
10.1109/ICPP.2009.60