Paper Info
Title
Evaluating the Performance Limitations of MPMD Communication
Abstract
The MPMD approach for parallel computing is attractive for programmers who seek fast development cycles, high code re-use, and modular programming, or whose applications exhibit irregular computation loads and communication patterns. RPC is widely adopted as the communication abstraction for crossing address space boundaries. However, the communication overheads of existing RPC-based systems are usually an order of magnitude higher than those found in highly tuned SPMD systems. This problem has thus far limited the appeal of high-level programming languages based on MPMD models in the parallel computing community. This paper investigates the fundamental limitations of MPMD communication using a case study of two parallel programming languages, Compositional C++ (CC++) and Split-C, that provide support for a global name space. To establish a common comparison basis, our implementation of CC++ was developed to use MRPC, a RPC system optimized for MPMD parallel computing and based on Active Messages. Basic RPC performance in CC++ is within a factor of two from those of Split-C and other messaging layers. CC++ applications perform within a factor of two to six from comparable Split-C versions, which represent an order of magnitude improvement over previous CC++ implementations. The results suggest that RPC-based communication can be used effectively in many high-performance MPMD parallel applications.
Year
DOI
Venue
1997
10.1109/SC.1997.10040
SC
Keywords
Field
DocType
mpmd approach,rpc-based communication,mpmd parallel computing,high-performance parallel application,communication overhead,high-performance mpmd parallel application,communication performance,communication pattern,mpmd model,performance limitations,communication abstraction,parallel computing community,performance limitation,mpmd communication,parallel computing,technical report,parallel computer,parallel programming,parallel programming language,gang scheduling,computer languages,multiprogramming,application software,parallel processing,computer science,concurrent computing,programming language,scheduling,computer applications
Address space,SPMD,Scheduling (computing),Computer science,Parallel computing,Gang scheduling,Modular programming,Computer multitasking,Concurrent computing,Application software,Distributed computing
Conference
ISBN
Citations 
PageRank 
0-89791-985-8
2
0.36
References 
Authors
23
4
Name
Order
Citations
PageRank
Chi-chao Chang141431.02
Grzegorz Czajkowski2190699.19
Thorsten von Eicken3841108.61
Carl Kesselman4128601648.67