DocumentCode :
3221692
Title :
Interprocedural distribution assignment placement: more than just enhancing intraprocedural placing techniques
Author :
Knoop, Jens ; Mehofer, Eduard
Author_Institution :
Fakultat fur Math. und Inf., Passau Univ., Germany
fYear :
1997
fDate :
10-14 Nov 1997
Firstpage :
26
Lastpage :
37
Abstract :
Avoiding unnecessary remappings at run-time by means of a strategic distribution assignment placement (DAP) is a major means for improving the run-time efficiency of data-parallel programs on distributed-memory architectures. In Proc. Euro-Par ´97, pp. 364-73 (1997), we presented a novel and aggressive intraprocedural algorithm achieving this by eliminating partially redundant and partially dead distribution assignments. In this paper, we show how to enhance this approach interprocedurally. Surprisingly at first sight, it turns out that a straightforward adaption of the intraprocedural approach fails because central properties being valid for the intraprocedural case do not carry over to the interprocedural one, revealing severe anomalies. After discussing the essential differences and analogies of DAP in the interprocedural and interprocedural cases, we show how to overcome these anomalies in order to arrive at a powerful and flexible approach for interprocedural DAP (IDAP). As in the interprocedural case, we get a hierarchy of IDAP algorithms of varying power and efficiency supporting user-customized solutions. First practical experiences underline its importance and effectivity
Keywords :
FORTRAN; data flow analysis; distributed memory systems; parallel algorithms; parallel programming; redundancy; subroutines; High Performance Fortran; algorithm hierarchy; anomalies; data-parallel programs; distributed-memory architectures; dynamic data redistribution; interprocedural data-flow analysis; interprocedural distribution assignment placement; intraprocedural placing techniques; optimization; partially dead distribution assignment elimination; partially redundant distribution assignment elimination; run-time efficiency; unnecessary run-time remappings; user-customized solutions; Algorithms; Digital audio players; Libraries; Memory architecture; Radio access networks; Runtime; Testing; Time measurement;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Parallel Architectures and Compilation Techniques., 1997. Proceedings., 1997 International Conference on
Conference_Location :
San Francisco, CA
Print_ISBN :
0-8186-8090-3
Type :
conf
DOI :
10.1109/PACT.1997.644001
Filename :
644001
Link To Document :
بازگشت