Hierarchical Dataflow Model for Efficient Programming of Clustered Manycore Processors

Julien Hascoët (1, 2) , Karol Desnos (1) , Jean-François Nezan (1) , Benoît Dupont de Dinechin (2)

1 IETR - Institut d'Électronique et des Technologies du numéRique
2 Kalray

Programming Multiprocessor Systems-on-Chips (MPSoCs) with hundreds of heterogeneous Processing Elements (PEs), complex memory architectures, and Networks-on-Chips (NoCs) remains a challenge for embedded system designers. Dataflow Models of Computation (MoCs) are increasingly used for developing parallel applications as their high-level of abstraction eases the automation of mapping, task scheduling and memory allocation onto MPSoCs. This paper introduces a technique for deploying hierarchical dataflow graphs efficiently onto MPSoC. The proposed technique exploits different granularity of dataflow parallelism to generate both NoC-based communications and nested OpenMP loops. Deployment of an image processing application on a many-core MPSoC results in speedups of up to 58.7 compared to the sequential execution.

Domaines: Informatique [cs]

Julien Hascoët, Karol Desnos, Jean-François Nezan, Benoît Dupont de Dinechin. Hierarchical Dataflow Model for Efficient Programming of Clustered Manycore Processors. 28th Annual IEEE International Conference on Application-specific Systems, Architectures and Processors (ASAP 2017), Jul 2017, Seattle, WA, United States. ⟨hal-01564019⟩