Reducing Memory Requirements of Stream Programs by Graph Transformations
BARTHOU, Denis
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Efficient runtime systems for parallel architectures [RUNTIME]
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Efficient runtime systems for parallel architectures [RUNTIME]
BARTHOU, Denis
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Efficient runtime systems for parallel architectures [RUNTIME]
< Reduce
Laboratoire Bordelais de Recherche en Informatique [LaBRI]
Efficient runtime systems for parallel architectures [RUNTIME]
Language
en
Communication dans un congrès
This item was published in
International IEEE Conference on High Performance Computing and Simulation, International IEEE Conference on High Performance Computing and Simulation, International IEEE Conference on High Performance Computing and Simulation, 2011-06-28, Caen. 2010-06-28p. p171-180
English Abstract
Stream languages explicitly describe fork-join parallelism and pipelines, offering a powerful programming model for many-core Multi-Processor Systems on Chip (MPSoC). In an embedded resource-constrained system, adapting ...Read more >
Stream languages explicitly describe fork-join parallelism and pipelines, offering a powerful programming model for many-core Multi-Processor Systems on Chip (MPSoC). In an embedded resource-constrained system, adapting stream programs to fit memory requirements is particularly important. In this paper we present a new approach to re- duce the memory footprint required to run stream programs on MPSoC. Through an exploration of equivalent program variants, the method selects parallel code minimizing mem- ory consumption. For large program instances, a heuristic accelerating the exploration phase is proposed and evalu- ated. We demonstrate the interest of our method on a panel of ten significant benchmarks. Using a multi-core modulo scheduling technique, our approach lowers considerably the minimal amount of memory required to run seven of these benchmarks while preserving throughput.Read less <
English Keywords
Graph Transformations
Stream Languages
Data Flow
Mem- ory
Graph Transformations.
Origin
Hal imported