System level performance analysis – the SymTA/S approach

Abstract

SymTA/S is a system-level performance and timing analysis approach based on formal scheduling analysis techniques and symbolic simulation. The tool supports heterogeneous architectures, complex task dependencies and context aware analysis. It determines system-level performance data such as end-to-end latencies, bus and processor utilization, and worst-case scheduling scenarios. SymTA/S furthermore combines optimization algorithms with system sensitivity analysis for rapid design space exploration. This paper gives an overview of the current research interests in the SymTA/S project.

MoVES — A framework for modelling and verifying embedded systems

“…Other tools supporting schedulability analysis of multicore systems are SymTA [5] and real-time-calculus [10]. Both systems are based on arrival-and service curves from Network Calculus [4] and analysis is performed using overapproximations.…”

Modeling of Ethernet AVB Networks for Worst-Case Timing Analysis

“…an upper bound on time required to process n activations of the task. This allows the computation of an output event model describing the minimum and maximum numbers of tasks completion in a specific interval as described in Henia et al (2005) and Schliecker et al (2008). The output event models are then forwarded as input event models of the dependent tasks, which are then analyzed again using the updated event models.…”

“…To analyze the worst-case timing behavior of Ethernet AVB, we use the Compositional Performance Analysis (CPA) approach as implemented by the tool SymTA/S (see Richter et al (2003); Henia et al (2005); Schliecker et al (2009)), which is similar to Real-Time Calculus (Thiele et al (2000)). This formal analysis assumes a set of tasks processed by communication or processing resources.…”

“…This procedure is iterated until a fixed point (stable event models) is reached or a timing constraint (e.g. maximum path latency) is violated (see Henia et al (2005)). To break cyclic analysis dependencies, initial event models for all tasks are derived from the external input event model of each task chain.…”

“…To provide accurate upper bounds on latency and hence formal guarantees, a formal compositional performance analysis (CPA) approach can be used which is based on the schedulability and timing analysis of distributed computing systems as described in e.g. Henia et al (2005). To this end, this paper presents a model transformation to transform an Ethernet AVB model into a timing analysis model.…”