This thesis contributes to the field of performance analysis in High Performance Computing with new concepts for in-memory event tracing. Event tracing records runtime events of an application and stores each with a precise time stamp and further relevant metrics. The high resolution and detailed information allows an in-depth analysis of the dynamic program behavior, interactions in parallel applications, and potential performance issues. For long-running and large-scale parallel applications, event-based tracing faces three challenges, yet unsolved: the number of resulting trace files limits scalability, the huge amounts of collected data overwhelm file systems and analysis capabilities, and the measurement bias, in particular, due to intermediate memory buffer flushes prevents a correct analysis. This thesis proposes concepts for an in-memory event tracing workflow. These concepts include new enhanced encoding techniques to increase memory efficiency and novel strategies for runtime event reduction to dynamically adapt trace size during runtime. An in-memory event tracing workflow based on these concepts meets all three challenges: First, it not only overcomes the scalability limitations due to the number of resulting trace files but eliminates the overhead of file system interaction altogether. Second, the enhanced encoding techniques and event reduction lead to remarkable smaller trace sizes. Finally, an in-memory event tracing workflow completely avoids intermediate memory buffer flushes, which minimizes measurement bias and allows a meaningful performance analysis. The concepts further include the Hierarchical Memory Buffer data structure, which incorporates a multi-dimensional, hierarchical ordering of events by common metrics, such as time stamp, calling context, event class, and function call duration. This hierarchical ordering allows a low-overhead event encoding, event reduction and event filtering, as well as new hierarchy-aided analysis requests. An experimental evaluation based on real-life applications and a detailed case study underline the capabilities of the concepts presented in this thesis. The new enhanced encoding techniques reduce memory allocation during runtime by a factor of 3.3 to 7.2, while at the same do not introduce any additional overhead. Furthermore, the combined concepts including the enhanced encoding techniques, event reduction, and a new filter based on function duration within the Hierarchical Memory Buffer remarkably reduce the resulting trace size up to three orders of magnitude and keep an entire measurement within a single fixed-size memory buffer, while still providing a coarse but meaningful analysis of the application. This thesis includes a discussion of the state-of-the-art and related work, a detailed presentation of the enhanced encoding techniques, the event reduction strategies, the Hierarchical Memory Buffer data structure, and a extensive experimental evaluation of all concepts.

The efficient parallel execution of scientific applications is a key challenge in high-performance computing (HPC). With growing parallelism and heterogeneity of compute resources as well as increasingly complex software, performance analysis has become an indispensable tool in the development and optimization of parallel programs. This thesis presents a framework for systematic performance analysis of scalable, heterogeneous applications. Based on event traces, it automatically detects the critical path and inefficiencies that result in waiting or idle time, e.g. due to load imbalances between parallel execution streams. As a prerequisite for the analysis of heterogeneous programs, this thesis specifies inefficiency patterns for computation offloading. Furthermore, an essential contribution was made to the development of tool interfaces for OpenACC and OpenMP, which enable a portable data acquisition and a subsequent analysis for programs with offload directives. At present, these interfaces are already part of the latest OpenACC and OpenMP API specification. The aforementioned work, existing preliminary work, and established analysis methods are combined into a generic analysis process, which can be applied across programming models. Based on the detection of wait or idle states, which can propagate over several levels of parallelism, the analysis identifies wasted computing resources and their root cause as well as the critical-path share for each program region. Thus, it determines the influence of program regions on the load balancing between execution streams and the program runtime. The analysis results include a summary of the detected inefficiency patterns and a program trace, enhanced with information about wait states, their cause, and the critical path. In addition, a ranking, based on the amount of waiting time a program region caused on the critical path, highlights program regions that are relevant for program optimization. The scalability of the proposed performance analysis and its implementation is demonstrated using High-Performance Linpack (HPL), while the analysis results are validated with synthetic programs. A scientific application that uses MPI, OpenMP, and CUDA simultaneously is investigated in order to show the applicability of the analysis.

No description available.

No description available.

Systems for high performance computing are getting increasingly complex. On the one hand, the number of processors is increasing. On the other hand, the individual processors are getting more and more powerful. In recent years, the latter is to a large extent achieved by increasing the number of cores per processor. Unfortunately, scientific applications often fail to fully utilize the available computational performance. Therefore, performance analysis tools that help to localize and fix performance problems are indispensable. Large scale systems for high performance computing typically consist of multiple compute nodes that are connected via network. Performance analysis tools that analyze performance problems that arise from using multiple nodes are readily available. However, the increasing number of cores per processor that can be observed within the last decade represents a major change in the node architecture. Therefore, this work concentrates on the analysis of the node performance. The goal of this thesis is to improve the understanding of the achieved application performance on existing hardware. It can be observed that the scaling of parallel applications on multi-core processors differs significantly from the scaling on multiple processors. Therefore, the properties of shared resources in contemporary multi-core processors as well as remote accesses in multi-processor systems are investigated and their respective impact on the application performance is analyzed. As a first step, a comprehensive suite of highly optimized micro-benchmarks is developed. These benchmarks are able to determine the performance of memory accesses depending on the location and coherence state of the data. They are used to perform an in-depth analysis of the characteristics of memory accesses in contemporary multi-processor systems, which identifies potential bottlenecks. However, in order to localize performance problems, it also has to be determined to which extend the application performance is limited by certain resources. Therefore, a methodology to derive metrics for the utilization of individual components in the memory hierarchy as well as waiting times caused by memory accesses is developed in the second step. The approach is based on hardware performance counters, which record the number of certain hardware events. The developed micro-benchmarks are used to selectively stress individual components, which can be used to identify the events that provide a reasonable assessment for the utilization of the respective component and the amount of time that is spent waiting for memory accesses to complete. Finally, the knowledge gained from this process is used to implement a visualization of memory related performance issues in existing performance analysis tools. The results of the micro-benchmarks reveal that the increasing number of cores per processor and the usage of multiple processors per node leads to complex systems with vastly different performance characteristics of memory accesses depending on the location of the accessed data. Furthermore, it can be observed that the aggregated throughput of shared resources in multi-core processors does not necessarily scale linearly with the number of cores that access them concurrently, which limits the scalability of parallel applications. It is shown that the proposed methodology for the identification of meaningful hardware performance counters yields useful metrics for the localization of memory related performance limitations.

This thesis presents how performance data from hardware accelerators can be included in event logs. It extends the capabilities of trace-based performance analysis to also monitor and record data from this novel parallelization layer. The increasing awareness to power consumption of computing devices has led to an interest in hybrid computing architectures as well. High-end computers, workstations, and mobile devices start to employ hardware accelerators to offload computationally intense and parallel tasks, while at the same time retaining a highly efficient scalar compute unit for non-parallel tasks. This execution pattern is typically asynchronous so that the scalar unit can resume other work while the hardware accelerator is busy. Performance analysis tools provided by the hardware accelerator vendors cover the situation of one host using one device very well. Yet, they do not address the needs of the high performance computing community. This thesis investigates ways to extend existing methods for recording events from highly parallel applications to also cover scenarios in which hardware accelerators aid these applications. After introducing a generic approach that is suitable for any API based acceleration paradigm, the thesis derives a suggestion for a generic performance API for hardware accelerators and its implementation with NVIDIA CUPTI. In a next step the visualization of event logs containing data from execution streams on different levels of parallelism is discussed. In order to overcome the limitations of classic performance profiles and timeline displays, a graph-based visualization using Parallel Performance Flow Graphs (PPFGs) is introduced. This novel technical approach is using program states in order to display similarities and differences between the potentially very large number of event streams and, thus, enables a fast way to spot load imbalances. The thesis concludes with the in-depth analysis of a case-study of PIConGPU---a highly parallel, multi-hybrid plasma physics simulation---that benefited greatly from the developed performance analysis methods. / Diese Dissertation zeigt, wie der Ablauf von Anwendungsteilen, die auf Hardwarebeschleuniger ausgelagert wurden, als Programmspur mit aufgezeichnet werden kann. Damit wird die bekannte Technik der Leistungsanalyse von Anwendungen mittels Programmspuren so erweitert, dass auch diese neue Parallelitätsebene mit erfasst wird. Die Beschränkungen von Computersystemen bezüglich der elektrischen Leistungsaufnahme hat zu einer steigenden Anzahl von hybriden Computerarchitekturen geführt. Sowohl Hochleistungsrechner, aber auch Arbeitsplatzcomputer und mobile Endgeräte nutzen heute Hardwarebeschleuniger um rechenintensive, parallele Programmteile auszulagern und so den skalaren Hauptprozessor zu entlasten und nur für nicht parallele Programmteile zu verwenden. Dieses Ausführungsschema ist typischerweise asynchron: der Skalarprozessor kann, während der Hardwarebeschleuniger rechnet, selbst weiterarbeiten. Die Leistungsanalyse-Werkzeuge der Hersteller von Hardwarebeschleunigern decken den Standardfall (ein Host-System mit einem Hardwarebeschleuniger) sehr gut ab, scheitern aber an einer Unterstützung von hochparallelen Rechnersystemen. Die vorliegende Dissertation untersucht, in wie weit auch multi-hybride Anwendungen die Aktivität von Hardwarebeschleunigern aufzeichnen können. Dazu wird die vorhandene Methode zur Erzeugung von Programmspuren für hochparallele Anwendungen entsprechend erweitert. In dieser Untersuchung wird zuerst eine allgemeine Methodik entwickelt, mit der sich für jede API-gestützte Hardwarebeschleunigung eine Programmspur erstellen lässt. Darauf aufbauend wird eine eigene Programmierschnittstelle entwickelt, die es ermöglicht weitere leistungsrelevante Daten aufzuzeichnen. Die Umsetzung dieser Schnittstelle wird am Beispiel von NVIDIA CUPTI darstellt. Ein weiterer Teil der Arbeit beschäftigt sich mit der Darstellung von Programmspuren, welche Aufzeichnungen von den unterschiedlichen Parallelitätsebenen enthalten. Um die Einschränkungen klassischer Leistungsprofile oder Zeitachsendarstellungen zu überwinden, wird mit den parallelen Programmablaufgraphen (PPFGs) eine neue graphenbasisierte Darstellungsform eingeführt. Dieser neuartige Ansatz zeigt eine Programmspur als eine Folge von Programmzuständen mit gemeinsamen und unterchiedlichen Abläufen. So können divergierendes Programmverhalten und Lastimbalancen deutlich einfacher lokalisiert werden. Die Arbeit schließt mit der detaillierten Analyse von PIConGPU -- einer multi-hybriden Simulation aus der Plasmaphysik --, die in großem Maße von den in dieser Arbeit entwickelten Analysemöglichkeiten profiert hat.

The assessment of performance for PEM Fuel Cells (PEMFC) at the stack, Single Repeating Unit (SRU), and Membrane Electrode Assembly (MEA) level is dominated by the evaluation of polarization curves. However, polarization curves do not provide adequate detail as to the origin of the inefficiencies of the fuel cell performance and information on these sources of origin are critical to understand and address topics such as material selection, optimal operating conditions, and overall robust and reliable cell and stack design characteristics. To the purpose of understanding the origin of the inefficiencies underlying the fuel cell polarization curve a series of additional experimental and analysis techniques must be applied and from the resultant data the origin of the inefficiencies can then be assigned to kinetic, ohmic, and mass transport loss categorizations. Further, through a combination of the diagnostic methods further resolution can be implied down to the contribution of the individual components to the relative voltage loss categories. In this topic, a methodology will be presented and discussed that achieves and demonstrates this process.

This thesis contributes to the field of performance analysis in High Performance Computing with new concepts for in-memory event tracing. Event tracing records runtime events of an application and stores each with a precise time stamp and further relevant metrics. The high resolution and detailed information allows an in-depth analysis of the dynamic program behavior, interactions in parallel applications, and potential performance issues. For long-running and large-scale parallel applications, event-based tracing faces three challenges, yet unsolved: the number of resulting trace files limits scalability, the huge amounts of collected data overwhelm file systems and analysis capabilities, and the measurement bias, in particular, due to intermediate memory buffer flushes prevents a correct analysis. This thesis proposes concepts for an in-memory event tracing workflow. These concepts include new enhanced encoding techniques to increase memory efficiency and novel strategies for runtime event reduction to dynamically adapt trace size during runtime. An in-memory event tracing workflow based on these concepts meets all three challenges: First, it not only overcomes the scalability limitations due to the number of resulting trace files but eliminates the overhead of file system interaction altogether. Second, the enhanced encoding techniques and event reduction lead to remarkable smaller trace sizes. Finally, an in-memory event tracing workflow completely avoids intermediate memory buffer flushes, which minimizes measurement bias and allows a meaningful performance analysis. The concepts further include the Hierarchical Memory Buffer data structure, which incorporates a multi-dimensional, hierarchical ordering of events by common metrics, such as time stamp, calling context, event class, and function call duration. This hierarchical ordering allows a low-overhead event encoding, event reduction and event filtering, as well as new hierarchy-aided analysis requests. An experimental evaluation based on real-life applications and a detailed case study underline the capabilities of the concepts presented in this thesis. The new enhanced encoding techniques reduce memory allocation during runtime by a factor of 3.3 to 7.2, while at the same do not introduce any additional overhead. Furthermore, the combined concepts including the enhanced encoding techniques, event reduction, and a new filter based on function duration within the Hierarchical Memory Buffer remarkably reduce the resulting trace size up to three orders of magnitude and keep an entire measurement within a single fixed-size memory buffer, while still providing a coarse but meaningful analysis of the application. This thesis includes a discussion of the state-of-the-art and related work, a detailed presentation of the enhanced encoding techniques, the event reduction strategies, the Hierarchical Memory Buffer data structure, and a extensive experimental evaluation of all concepts.

Die Dissertation behandelt methodische und algorithmische Aspekte der Spektralelementemethode zur räumlichen Diskretisierung partieller Differentialgleichungen. Die Weiterentwicklung einer symmetriebasierten Faktorisierung ermöglicht effiziente Operatoren für Tetraederelemente. Auf Grundlage einer umfassenden Leistungsanalyse werden Engpässe in der Implementierung der Operatoren identifiziert und durch algorithmische Modifikationen der Methode eliminiert.

The use of accelerators in heterogeneous systems is an established approach in designing petascale applications. Today, Compute Unified Device Architecture (CUDA) offers a rich programming interface for GPU accelerators but requires developers to incorporate several layers of parallelism on both the CPU and the GPU. From this increasing program complexity emerges the need for sophisticated performance tools. This work contributes by analyzing hybrid MPICUDA programs for properties based on wait states, such as the critical path, a metric proven to identify application bottlenecks effectively. We developed a tool to construct a dependency graph based on an execution trace and the inherent dependencies of the programming models CUDA and Message Passing Interface (MPI). Thereafter, it detects wait states and attributes blame to responsible activities. Together with the property of being on the critical path, we can identify activities that are most viable for optimization. To evaluate the global impact of optimizations to critical activities, we predict the program execution using a graph-based performance projection. The developed approach has been demonstrated with suitable examples to be both scalable and correct. Furthermore, we establish a new categorization of CUDA inefficiency patterns ensuing from the dependencies between CUDA activities.