Abstract:The analysis of execution traces can reveal important information about the behavioral aspects of complex software systems, hence reducing the time and effort it takes to understand and maintain them. Traces, however, tend to be considerably large which hinders their effective analysis. Existing traces analysis tools rely on some sort of visualization techniques to help software engineers make sense of trace content. Many of these techniques have been studied and found to be limited in many ways. In this paper… Show more
“…Pirzadeh and Hamou-Lhadj presented a novel phase detection approach that they called trace segmentation and which was inspired by the way the human perception system groups lines and dots into shapes and objects [23,30]. They have developed several methods that could automatically group trace events into dense elements that formed computational phases.…”
Section: Related Workmentioning
confidence: 99%
“…For example, Process 7 will send to and receive from processes 2, 3,4,6,8,10,11,12,18,19,20,21,22,23,24,26,27, and 28.…”
Understanding the behaviour of High Performance Computing (HPC) systems is a challenging task due to the large number of processes they involve as well as the complex interactions among these processes. In this paper, we present a novel approach that aims to simplify the analysis of large execution traces generated from HPC applications. We achieve this through a technique that allows semi-automatic extraction of execution phases from large traces. These phases, which characterize the main computations of the traced scenario, can be used by software engineers to browse the content of a trace at different levels of abstraction. Our approach is based on the application of information theory principles to the analysis of sequences of communication patterns found in HPC traces. The results of the proposed approach when applied to traces of a large HPC industrial system demonstrate its effectiveness in identifying the main program phases and their corresponding sub-phases.
“…Pirzadeh and Hamou-Lhadj presented a novel phase detection approach that they called trace segmentation and which was inspired by the way the human perception system groups lines and dots into shapes and objects [23,30]. They have developed several methods that could automatically group trace events into dense elements that formed computational phases.…”
Section: Related Workmentioning
confidence: 99%
“…For example, Process 7 will send to and receive from processes 2, 3,4,6,8,10,11,12,18,19,20,21,22,23,24,26,27, and 28.…”
Understanding the behaviour of High Performance Computing (HPC) systems is a challenging task due to the large number of processes they involve as well as the complex interactions among these processes. In this paper, we present a novel approach that aims to simplify the analysis of large execution traces generated from HPC applications. We achieve this through a technique that allows semi-automatic extraction of execution phases from large traces. These phases, which characterize the main computations of the traced scenario, can be used by software engineers to browse the content of a trace at different levels of abstraction. Our approach is based on the application of information theory principles to the analysis of sequences of communication patterns found in HPC traces. The results of the proposed approach when applied to traces of a large HPC industrial system demonstrate its effectiveness in identifying the main program phases and their corresponding sub-phases.
“…Pirzadeh et al in [2] use analysis of execution traces to understand behavioural aspects of complex software systems. They can divide the content of a trace into meaningful trace segments called execution phases, and their slicing is done using Gestalt laws.…”
Section: Related Workmentioning
confidence: 99%
“…However, the programmer has to go through a series of pages in order to find a specific information. Our work divides an input trace into relevant blocks, mutually independents, without any information on some execution part, unlike [2,10]. The visualization step becomes now easier to manage through information grouping, and the analysis is improved.…”
Section: Related Workmentioning
confidence: 99%
“…These techniques can obtain a reduced execution trace but not always representative of the entire trace [9]. Pirzadeh et al introduced in [2] that the general consensus in the trace analysis community is to emphasise the work towards effective trace abstraction techniques, such as [10]. For reducing the volume of events in a trace to be analysed by the programmer, we propose to abstract series of lowlevel events as blocks that are meaningful to the programmer, and then to further abstract the trace as sequences of blocks.…”
ISBN: 978-1-4673-5164-5 - http://icdm2012.ua.ac.be/International audienceThe analysis of multimedia application traces can reveal important information to enhance program comprehension. However traces can be very large, which hinders their effective exploitation. In this paper, we study the problem of finding a \textit{k-golden} set of blocks that best characterize data. Sequential pattern mining can help to automatically discover the blocks, and we called \textit{k-golden set}, a set of $k$ blocks that maximally covers the trace. These kind of blocks can simplify the exploration of large traces by allowing programmers to see an abstraction instead of low-level events. We propose an approach for mining golden blocks and finding coverage of frames. The experiments carried out on video and audio application decoding show very promising results
Abstract-Identifying concepts in execution traces is a task often necessary to support program comprehension or maintenance activities. Several approaches-static, dynamic or hybrid-have been proposed to identify cohesive, meaningful sequence of methods in execution traces. However, none of the proposed approaches is able to label such segments and to identify relations between segments of the same trace.This paper present SCAN (Segment Concept AssigNer) an approach to assign labels to sequences of methods in execution traces, and to identify relations between such segments. SCAN uses information retrieval methods and formal concept analysis to produce sets of words helping the developer to understand the concept implemented by a segment. Specifically, formal concept analysis allows SCAN to discover commonalities between segments in different trace areas, as well as terms more specific to a given segment and high level relations between segments.The paper describes SCAN along with a preliminary manual validation-upon execution traces collected from usage scenarios of JHotDraw and ArgoUML-of SCAN accuracy in assigning labels representative of concepts implemented by trace segments.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.