Title :
Optimizing bulk synchronous parallel ML
Author :
Loulergue, Frédéric
Author_Institution :
Lab. of Algorithms, Complexity & Logic, Univ. Paris Val-de-Marne, Creteil, France
Abstract :
Bulk synchronous parallel ML is a functional parallel language based on the bulk synchronous parallelism model of computation. Deadlocks are avoided and programs are deterministic. The performance of programs can be accurately predicted. This paper addresses the optimization of BSML programs through the compilation of BSML primitives to Objective Caml code with calls to a low level library rather than their direct implementation as a high level library.
Keywords :
ML language; parallel programming; software libraries; software performance evaluation; bulk synchronous parallel ML; bulk synchronous parallelism model; deadlocks; functional parallel language; low level library; objective Caml code; Computer networks; Concurrent computing; Costs; Distributed computing; Laboratories; Libraries; Logic; Parallel languages; Parallel processing; System recovery;
Conference_Titel :
Software Engineering, Artificial Intelligence, Networking and Parallel/Distributed Computing, 2005 and First ACIS International Workshop on Self-Assembling Wireless Networks. SNPD/SAWN 2005. Sixth International Conference on
Print_ISBN :
0-7695-2294-7
DOI :
10.1109/SNPD-SAWN.2005.56
