Partial Expansion Graphs: Exposing Parallelism and Dynamic Scheduling Opportunities for DSP Applications

Abstract

Emerging Digital Signal Processing (DSP) algorithms and wireless communications protocols require dynamic adaptation and online reconfiguration for the implemented systems at runtime. In this paper, we introduce the concept of Partial Expansion Graphs (PEGs) as an implementation model and associated class of scheduling strategies. PEGs are designed to help realize DSP systems in terms of forms and granularities of parallelism that are well matched to the given applications and targeted platforms. PEGs also facilitate derivation of both static and dynamic scheduling techniques, depending on the amount of variability in task execution times and other operating conditions. We show how to implement efficient PEG-based scheduling methods using real time operating systems, and to re-use pre-optimized libraries of DSP components within such implementations. Empirical results show that the PEG strategy can 1) achieve significant speedups on a state of the art multicore signal processor platform for static dataflow applications with predictable execution times, and 2) exceed classical scheduling speedups for applications having execution times that can vary dynamically. This ability to handle variable execution times is especially useful as DSP applications and platforms increase in complexity and adaptive behavior, thereby reducing execution time predictability.

Partial expansion of dataflow graphs for resource-aware scheduling of multicore signal processing systems

“…For the three versions of the pseudo-communication benchmark, the same speedups that were obtained manually (see [3]) 3 (a) Digital receiver benchmark.…”

“…We encourage the reader to review [3] for further background about the PEG model, and for detailed examples that help to present key principles of the model in a concrete manner.…”

“…In our previous work [3], we explored the realization of the PEG model assuming that we already have a PEG scheduling solution consisting of given sets of data parallel actor expansions, and given mappings of actors to subsets of processors. In this paper, we build on the PEG model in three ways.…”

“…In [3], we developed Partial Expansion Graphs (PEGs) as an implementation model and associated class of scheduling strategies for graphs represented using SDF. PEGs are designed to help realize DSP systems in terms of forms and granularity levels of parallelism that are well matched to the given applications and targeted platforms.…”

Numerical Representation of Directed Acyclic Graphs for Efficient Dataflow Embedded Resource Allocation

“…There is a paradox between developing more expressive and more compact dataflow MoCs, and the fact that analysis methods often depend on the need of expanding expressive graphs into DAGs. Some works, however, try to take advantage of the expressiveness of the original MoC [8] or to limit the expansion of graphs and accelerate analysis [23].…”