Abstract | ||
---|---|---|
The pipeline is a simple and intuitive structure to speed up many problems. Novice parallel programmers are usually taught this structure early on. However, expert parallel programmers typically eschew using the pipeline in coarse-grained applications because it has three serious problems that make it difficult to implement efficiently. First, processors are idle when the pipeline is not full. Second, load balancing is crucial to obtaining good speedup. Third, it is difficult to incrementally incorporate more processors into an existing pipeline. Instead, experts recast the problem as a master/slave structure which does not suffer from these problems. This paper details a transformation that allows programs written in a pipeline style to execute using the master/slave structure. Parallel programmers can benefit from both the intuitive simplicity of the pipeline and the efficient execution of a master/slave structure. This is demonstrated by performance results from two applications. |
Year | DOI | Venue |
---|---|---|
2004 | 10.1109/HIPS.2004.1299186 | NINTH INTERNATIONAL WORKSHOP ON HIGH-LEVEL PARALLEL PROGRAMMING MODELS AND SUPPORTIVE ENVIRONMENTS, PROCEEDINGS |
Keywords | Field | DocType |
decoding,object oriented,load balancing,master slave,load balance,resource allocation,concurrent computing,computer science,object oriented programming,pipelines,parallel programming | Pipeline (computing),Graphics pipeline,Object-oriented programming,Computer science,Load balancing (computing),Idle,Parallel computing,Resource allocation,Pipeline (software),Distributed computing,Speedup | Conference |
Citations | PageRank | References |
17 | 1.31 | 7 |
Authors | ||
3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Steve Macdonald | 1 | 139 | 13.66 |
D. Szafron | 2 | 1579 | 210.88 |
Jonathan Schaeffer | 3 | 1929 | 220.77 |