Mitigating Software-Instrumentation Cache Effects in Measurement-Based Timing Analysis
Enrique Díaz, Jaume Abella, Enrico Mezzetti, Irune Agirre, Mikel Azkarate-Askasua, Tullio Vardanega, Francisco J. Cazorla
Abstract
Enrique Díaz, Jaume Abella, Enrico Mezzetti, Irune Agirre, Mikel Azkarate-Askasua, Tullio Vardanega, Francisco J. Cazorla
Abstract
Measurement-based timing analysis (MBTA) is often used to determine the timing behaviour of software programs embedded in safety-aware real-time systems deployed in various industrial domains including automotive and railway. MBTA methods rely on some form of instrumentation, either at hardware or software level, of the target program or fragments thereof to collect execution-time measurement data. A known drawback of software-level instrumentation is that instrumentation itself does affect the timing and functional behaviour of a program, resulting in the so-called probe effect: leaving the instrumentation code in the final executable can negatively affect average performance and could not be even admissible under stringent industrial qualification and certification standards; removing it before operation jeopardizes the results of timing analysis as the WCET estimates on the instrumented version of the program cannot be valid any more due, for example, to the timing effects incurred by different cache alignments. In this paper, we present a novel approach to mitigate the impact of instrumentation code on cache behaviour by reducing the instrumentation overhead while at the same time preserving and consolidating the results of timing analysis.
OpenAlex reports 5 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
Measurement-based timing analysis (MBTA) is often used to determine the timing behaviour of software programs embedded in safety-aware real-time systems deployed in various industrial domains including automotive and railway. MBTA methods rely on some form of instrumentation, either at hardware or software level, of the target program or fragments thereof to collect execution-time measurement data. A known drawback of software-level instrumentation is that instrumentation itself does affect the timing and functional behaviour of a program, resulting in the so-called probe effect: leaving the instrumentation code in the final executable can negatively affect average performance and could not be even admissible under stringent industrial qualification and certification standards; removing it before operation jeopardizes the results of timing analysis as the WCET estimates on the instrumented version of the program cannot be valid any more due, for example, to the timing effects incurred by different cache alignments. In this paper, we present a novel approach to mitigate the impact of instrumentation code on cache behaviour by reducing the instrumentation overhead while at the same time preserving and consolidating the results of timing analysis.
Key concepts: Instrumentation (computer programming), Executable, Computer science, Cache, Software, Embedded system, Static timing analysis, Overhead (engineering)