Spider: A Synchronous Parameterized and Interfaced Dataflow-based RTOS for multicore DSPS

Abstract

This paper introduces a novel Real-Time Operating System (RTOS) based on a parameterized dataflow Model of Computation (MoC). This RTOS, called Synchronous Parameterized and Interfaced Dataflow Embedded Runtime (SPiDER), aims at efficiently scheduling Parameterized and Interfaced Synchronous Dataflow (PiSDF) graphs on multicore architectures. It exploits features of PiSDF to locate locally static regions that exhibit predictable application behavior. This paper uses a multicore signal processing benchmark to demonstrate that the SPiDER runtime can exploit more parallelism than a conventional multicore task scheduler. By comparing experimental results of the SPiDER runtime on an 8-core Texas Instruments Keystone I Digital Signal Processor (DSP) with those obtained from the OpenMP framework, latency improvements of up to 26% are demonstrated.

Dataflow Modeling for Reconfigurable Signal Processing Systems

“…Spider (Synchronous Parameterized and Interfaced Dataflow Embedded Runtime) is a Real-Time Operating System (RTOS) whose purpose is to manage the execution of πSDF graphs on heterogeneous Multiprocessor Systems-on-Chips (MPSoCs) [29]. As presented in Figure 11, πSDF graphs and source code associated to actors, generally coded in C language, are designed by the application developer using the Parallel Real-time Embedded Executives Scheduling Method (Preesm) rapid prototyping framework [51].…”

“…Indeed, such runtime manager requires processor time to compute repetition vectors, to perform graph transformations, and to map and schedule actor firings. Nev-ertheless, as presented in [29,28], even with large reconfigurable graphs with several hundreds of actors, this overhead is largely compensated by the efficiency of the scheduling decisions, and generally outperforms dynamic deployment strategies with no global manager.…”

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

“…Our representation allows using DAG oriented analysis methods while maintaining the compactness and the expressiveness of the targeted dataflow MoC. We implemented our numerical modeling of DAGs in the Spider tool [13] on three different platforms ranging from a medium laptop to a low power embedded platform. Our experiments show a significant reduction of the overhead of the Spider embedded runtime both in terms of execution time and memory footprint of the runtime.…”

“…Then, our numerical representation of the DAG is presented in Section 4. Section 5 presents experimental results of the implementation of our contribution into the Spider tool [13] on signal processing applications. Finally, Section 6 concludes this paper.…”

Just-in-time scheduling techniques for multicore signal processing systems

“…The JIT-MS scheduling method is embedded in the Synchronous Parameterized and Interfaced Dataflow Embedded Runtime (SPIDER) [10] This paper is organized as follows: Section II and Section III present related research, providing the context of current work, Section IV details our proposed new JIT-MS scheduling method, and Section V presents our experimental results with JIT-MS scheduling.…”