A multi-DAG model for real-time parallel applications with conditional execution

Abstract

Owing to the current trends for higher performance and the ever growing availability of multiprocessors in the embedded computing (EC) domain, there is nowadays a strong push towards the parallelization of modern embedded applications. Several real-time task models have recently been proposed to capture different forms of parallelism. However, they do not deal explicitly with control flow information as they assume that all the threads of a parallel task must execute every time the task is activated. In contrast, in this paper, we present a multi-DAG model where each task is characterized by a set of execution flows, each of which represents a different execution path throughout the task code and is modeled as a DAG of sub-tasks. We propose a two-step solution that computes a single synchronous DAG of servers for a task modeled by a multi-DAG and show that these servers are able to supply every execution flow of that task with the required cpu-budget so that the task can execute entirely, irrespective of the execution flow taken at run-time, while satisfying its precedence constraints. As a result, each task can be modeled by its single DAG of servers, which facilitates in leveraging the existing single-DAG schedulability analyses techniques for analyzing the schedulability of parallel tasks with multiple execution flows. ABSTRACTOwing to the current trends for higher performance and the ever growing availability of multiprocessors in the embedded computing (EC) domain, there is nowadays a strong push towards the parallelization of modern embedded applications. Several real-time task models have recently been proposed to capture different forms of parallelism. However, they do not deal explicitly with control flow information as they assume that all the threads of a parallel task must execute every time the task is activated. In contrast, in this paper, we present a multi-DAG model where each task is characterized by a set of execution flows, each of which represents a different execution path throughout the task code and is modeled as a DAG of sub-tasks. We propose a two-step solution that computes a single synchronous DAG of servers for a task modeled by a multi-DAG and show that these servers are able to supply every execution flow of that task with the required cpu-budget so that the task can execute entirely, irrespective of the execution flow taken at run-time, while satisfying its precedence constraints. As a result, each task can be modeled by its single DAG of servers, which facilitates in leveraging the existing single-DAG schedulability analyses techniques for analyzing the schedulability of parallel tasks with multiple execution flows.

Schedulability Analysis of Conditional Parallel Task Graphs in Multicore Systems

“…We show the experimental comparison of our response-time analysis against the only two works in the literature that target the global scheduling of DAG tasks with conditional branches, i.e., [20] and [10]. The first work proposes a transformation of conditional DAG tasks scheduled with global FP into synchronous parallel tasks.…”

“…The first attempts at enriching a parallel task model with control-flow information were proposed in the context of uniprocessor systems to provide a more accurate characterization of the worst-case behavior of a task [5], [8], [17]. In a multicore setting, Fonseca et al [20] proposed the multi-DAG model, which represents a parallel task as a collection of DAGs, each representing a different execution flow. The authors proposed a method to combine such flows into a single synchronous parallel task that preserves the execution requirements and the precedence constraints of all the execution flows that can possibly occur at runtime, thus reducing the schedulability problem to a simpler problem for synchronous parallel tasks.…”

“…As noted by Fonseca et al [20], the problem introduced by conditional statements is particularly significant in the schedulability analysis of parallel tasks running on a multicore system, although explicitly modeling branching structures has been proven useful also for single-core systems to tighten the response time of the tasks [5], [8], [17]. The problem is exacerbated by the presence of multiple conditional statements characterized by branches with very different length.…”

An Execution Model for Fine-Grained Parallelism in Ada

“…Works that consider the general workload of a graph can still be used, by taking the maximum workload and critical path length among all the potential graphs (with a less optimal result), or approaches that use a single tighter worst-case graph can be considered [22], which allow one to reduce the complexity of the interference analysis.…”

Reducing Response-Time Bounds for DAG-Based Task Systems on Heterogeneous Multicore Platforms

“…Related work. The literature on real-time systems includes much work pertaining to the scheduling of DAG-based task systems on identical multiprocessor platforms [1,2,4,5,6,7,9,10,13,15,17,18,19,20,21,22,23,24]. However, we are not aware of any corresponding work directed at heterogeneous platforms.…”