Fault tolerant task mapping on many-core arrays

Bonney, Colin; Campos, Pedro; Dahir, Nizar; Tempesti, Gianluca

doi:10.1109/ssci.2016.7850174

Cited by 5 publications

(12 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One example of such a method is the evolutionary algorithm (EA), which presents a promising strategy for the exploration of very large solution space. In this paper, we use the EA-based mapping proposed in [5] for mapping a task graph into the SpiNNaker platform. The EA in [5] was applied on a many-core virtual platform.…”

Section: Heuristic Task Graph Mappingmentioning

confidence: 99%

“…In this paper, we use the EA-based mapping proposed in [5] for mapping a task graph into the SpiNNaker platform. The EA in [5] was applied on a many-core virtual platform. Hence, it did not take into account the pre-processing overhead of the communication channel of a real platform.…”

Section: Heuristic Task Graph Mappingmentioning

confidence: 99%

“…For mapping the entire task graph on SpiNNaker we opt to use a heuristic method. In this paper, we use the results of the EA proposed in [5], and propose an algorithm to further improve the overall performance of the mapped task graph. Our algorithm, which is called scattering-and-grouping, is basically a load-balancing strategy that can be applied on any existing map.…”

Section: Proposed Optimization Strategymentioning

confidence: 99%

“…In this paper, we evaluate the performance of our proposed improvement described in Section 3.3 using a 25-node graph shown in Fig.3, the same network used in [5].…”

Section: Experimental Set-upmentioning

confidence: 99%

“…In this paper we use static mappings generated by an evolutionary algorithm as proposed in [5]. The basic idea of our method is the balancing of data traffic across the SpiNNaker mesh.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Optimized task graph mapping on a many-core neuromorphic supercomputer

Sugiarto

Campos

Dahir

et al. 2017

2017 IEEE High Performance Extreme Computing Conference (HPEC)

Self Cite

View full text Add to dashboard Cite

Abstract-This paper presents an approach for improving the overall performance of a general purpose application running as a task graph on a many-core neuromorphic supercomputer. Our task graph framework is based on graceful degradation and amelioration paradigms that strive to achieve high reliability and performance by incorporating fault tolerance and task spawning features. The optimization is applied on an instance of the task graph by performing a soft load balancing on the data traffic between nodes in the graph. We implemented the framework and its optimization on SpiNNaker, a manycore neuromorphic platform containing a million ARM9 processing cores. We evaluate our method using several static mapping examples, where some of them were generated using an evolutionary algorithm. The experiment demonstrates that a performance improvement of up to 8.2% can be achieved when implementing our algorithm on a fully-utilized SpiNNaker communication infrastructure.

show abstract

Section: Heuristic Task Graph Mappingmentioning

confidence: 99%

Section: Heuristic Task Graph Mappingmentioning

confidence: 99%

Section: Proposed Optimization Strategymentioning

confidence: 99%

“…In this paper, we evaluate the performance of our proposed improvement described in Section 3.3 using a 25-node graph shown in Fig.3, the same network used in [5].…”

Section: Experimental Set-upmentioning

confidence: 99%

“…In this paper we use static mappings generated by an evolutionary algorithm as proposed in [5]. The basic idea of our method is the balancing of data traffic across the SpiNNaker mesh.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations