Name
Abstract | ||
|---|---|---|
Data-flow graphs, consisting of processes (actors) and communication channels, provide an efficient model of computation for analysis and implementation of highly parallel applications. We propose a novel algorithm for scheduling a large number of data-flow actors and also recursively expandable sub-graphs by encoding their state and dependencies as a token stream. The proposed execution model enables global resource sharing, dynamic instantiation of actors and provides lightweight lock-free synchronization. Hence, our approach is most useful for compute-intensive applications with frequently varying and unpredictable data rates. In addition, we present a balanced scheduling algorithm, which restricts the memory usage of dynamic and recursive data-flow graphs, while still maintaining enough concurrency to keep all execution units utilized. |
Year | DOI | Venue |
|---|---|---|
2014 | 10.1109/SBAC-PADW.2014.7 | Computer Architecture and High Performance Computing Workshop |
Keywords | Field | DocType |
data flow graphs,multiprocessing systems,rewriting systems,scheduling,synchronisation,compute-intensive applications,data-flow actor scheduling,dynamic actor instantiation,dynamic data-flow graph scheduling,frequently varying data rates,global resource sharing,lightweight lock-free synchronization,recursive data-flow graph scheduling,recursively expandable subgraphs,stream rewriting,unpredictable data rates | Fair-share scheduling,Computer science,Parallel computing,Gang scheduling,Real-time computing,Two-level scheduling,Rate-monotonic scheduling,Dynamic priority scheduling,Foreground-background,Earliest deadline first scheduling,Round-robin scheduling,Distributed computing | Conference |
Citations | PageRank | References |
1 | 0.36 | 21 |
Authors | ||
2 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Lars Middendorf | 1 | 16 | 4.07 |
| Christian Haubelt | 2 | 796 | 68.77 |