Logo des Repositoriums

Auflistung nach Schlagwort "Kieker"

1 - 4 von 4
  • Konferenzbeitrag
    Architecture Recovery from Fortran Code with Kieker
    (Softwaretechnik-Trends Band 43, Heft 1, 2023) Jung, Reiner; Schnoor, Henning; Gundlach, Sven; Hasselbring, Wilhelm
    Scientific models are software systems, which are key to understand and assess a range of challenges, such as climate change mitigation. However, such models are usually developed over decades. To support program comprehension for software maintenance and restructuring, we designed an architecture recovery process for Fortran-based scientific models utilizing Kieker 4 C to collect call traces at runtime. Furthermore, we derive structural information from the recovered architecture. In this paper, we present our analysis process and some results from analyzing three scientific models. Additionally, we describe how to use the information obtained by our analysis to identify possible optimizations of the scientific models.
  • Konferenzbeitrag
    Graph-Based Performance Analysis at System- and Application-Level
    (Softwaretechnik-Trends Band 40, Heft 3, 2020) Müller, Richard; Strempel, Tom
    The Kieker plugin for jQAssistant transforms monitored log data into graphs to support software engineers with performance analysis. In this paper, we describe how we have extended and improved this plugin to support performance analysis at system- and application-level and how we have evaluated its correctness and scalability using data from recent experiments. This is a first step to replicate complete experiments in the field of performance analysis using graphs.
  • Konferenzbeitrag
    Instrumenting Python with Kieker
    (Softwaretechnik-Trends Band 43, Heft 1, 2023) Simonov, Serafim; Düllmann, Thomas F.; Jung, Reiner; Gundlach, Sven
    Python has become a widely used programming language in big data, machine learning, and scientific modeling. In all these domains, performance is a key factor to success and requires the ability to understand the runtime behavior of software. Therefore, we ported Kieker monitoring to Python and evaluated different approaches to introduce probes into code. In this paper, we evaluate these approaches, show their benefits and limitations and provide a perfor mance evaluation of the Kieker 4 Python framework.
  • Konferenzbeitrag
    More is Less in Kieker? The Paradox of No Logging Being Slower Than Logging
    (Softwaretechnik-Trends Band 43, Heft 4, 2023) Reichelt, David Georg; Jung, Reiner; van Hoorn, André
    Understanding the sources of monitoring overhead is crucial for understanding the performance of a monitored application. The MooBench bench mark measures the monitoring overhead and its sources. MooBench assumes that benchmarking overhead emerges from the instrumentation, the data collection, and the writing of data. These three parts are measured through individual factorial experiments. We made the counter-intuitive observation that MooBench consistently and reproducibly reported higher overhead for Kieker and other monitoring frameworks when not writing data. Intuitively, writing should consume resources and therefore slow down (or, since is parallelized, at least not speed up) the monitoring. In this paper, we present an investigation of this problem in Kieker. We find that lock contention at Kieker’s writing queue causes the problem. Therefore, we propose to add a new queue that dumps all elements. Thereby, a realistic measurement of data collection without writing can be provided.