Technische Fakultät

Permanent URI for this collection

Browse

Recent Submissions

Now showing 1 - 20 of 4137
  • Doctoral thesis
    Open Access
    Software-based Memory-Consistency Mechanisms for Non-Coherent Many-Core Systems
    (2024) Langer, Tobias; Schröder-Preikschat, Wolfgang
    Multi- and manycore processors have become virtually indispensable. Some of today's manycore processors already have cores counts in the higher two digits numbers. As the performance hunger of modern applications further increases, this trend towards more parallel processors will likely continue. However, this development requires significant changes in computer architectures. As a consequence, modern processor designs already deviate considerably from that of classical SMP systems, as components such as bus systems do not scale with the core count, and memory connects increasingly become a bottleneck. One effect of these changes is that cache coherence mechanisms become increasingly complex. These systems already cause high engineering efforts, as they require long design times and are complicated and time intensive to verify. Furthermore, they become increasingly resource intensive as their chip area and energy consumption grow with parallelism. In the meantime, a different trend can be observed for system software. Cache coherence mechanisms have here been identified as a source of performance overheads. Therefore, many modern operating and runtime systems thus employ strategies to limit memory accesses to NUMA domains so that only the coherence mechanisms there are used. Cross-NUMA coherence, in return, is avoided, if possible. These software systems thus effectively avoid the coherent shared memory that the hardware level provides at great expense. However, shared memory programming still provides a more convenient programming model than explicit communication schemes, such as message passing. Considering this situation, this dissertation proposes that coherence mechanisms should be realized as part of the software layer instead of providing them in hardware. The coherence mechanism can thus adapt to the applications' requirements, making it far more flexible than hardware implementations can be. Therefore, this dissertation explores the design space for software-based coherence solutions. The proposed designs thereby build on techniques known from Virtual Shared Memory (VSM) systems to realize a memory consistency mechanism as an on-demand feature. This notion of memory consistency also meets the requirement for cache coherence. The results and contribution of this thesis are two variants of software-based consistency mechanisms. The first of which is the Software Consistency System (SCS), which follows the principles of classical VSM systems by leveraging the features of virtual memory management to create a consistent shared memory abstraction. The second variant is the Adaptive Software Consistency System (ASCS), a novel adaptive consistency mechanism. The ASCS builds upon the SCS and extends it with the ability to exploit memory-sharing information to switch adaptively between page and hardware caches at runtime. Finally, this thesis demonstrates the functionality and effectiveness of the consistency systems with a performance evaluation on a tile-based prototype system implemented in FPGA logic at the example of different shared memory workloads. The results show that the SCS and the ASCS provide good scalability for the parallel workloads and, thus, that they are feasible alternatives to existing hardware-based coherence mechanisms. Furthermore, the experiments show that for all but one evaluation scenario, the ASCS mechanism outperforms the SCS.
  • Doctoral thesis
    Open Access
    Quantifying Inner Source Software Development for Business Process Usage
    (2024) Buchner, Stefan; Riehle, Dirk
    Inner source refers to the application of open source development principles within corporate environments. Essentially, it involves embracing open collaboration and contribution without necessarily developing open source software. One fundamental concept borrowed from open source development is the practice of opening software projects to external contributions. In the realm of inner source, these contributions usually originate from developers engaged in other projects or working in different organizational units. This collaborative approach yields mutual benefits through joint software development. However, these inner source contributions often cross organizational and legal boundaries, significantly impacting a wide range of business processes. These effects extend beyond software development and affect strategic and operational functions across various departments, including management, accounting, and even taxation. Currently, these processes are not harmonized with the inner source way of intellectual property flow within the company. This misalignment can lead to mismanagement and even profit shifting. The primary objective of this dissertation is to address these challenges by making three consecutive contributions: First, we conducted a systematic literature review, combined with thematic analysis, to assess the current state of inner source measurement and its impact on various business processes. This qualitative data analysis categorizes a range of approaches and evaluates their suitability for application within the inner source domain. Second, building on the insights derived from our systematic literature review, we developed an inner source research model. This model serves as a foundation, providing researchers with a unified framework for advancing toward more precise and comprehensive inner source measurement. It, in turn, enables the development of future tools and techniques and facilitates the creation of metrics tailored to inner source management, aligning existing business processes with the inner source paradigm. Third, we have implemented a first algorithm for the financial assessment of inner source contributions. This algorithm estimates the time invested in code contributions by individual contributors or departments. These estimations, in turn, enable cost calculations that can be applied to business processes affected by inner source, such as tax contribution assessments. Lastly, we consolidate individual contributions, clarify thematic connections, integrate outside research, and explore future prospects. In summary, this dissertation lays the theoretical groundwork for the financial evaluation of inner source contributions. It offers guidance to researchers on conducting measurement-related inner source research and provides practitioners with insights into the development of inner source metrics and tools. Furthermore, this dissertation introduces an initial algorithm, practically evaluated for its usability in conducting cost calculations within the inner source domain.
  • Doctoral thesis
    Open Access
    On Aspects of Cardiac and Artificial Muscle Modelling – Insights into Orthotropic Tissue Structure and Dielectric Elastomer Actuators
    Schriftenreihe Technische Dynamik : 10, (2024) Holz, David; Leyendecker, Sigrid
    The heart is a fascinating organ whose seemingly simple function of mechanically pumping blood throughout the body is ensured by a complex interaction of mechanical, electrical, chemical and biological mechanisms. However, complex systems inherently bear the risk of disruptions, causing a variety of cardiovascular diseases (CVD) such as myocardial infarction, heart failure, arrhythmia, and hypertension. The field of biomechanics is playing a key role in the fundamental research concerning CVDs. In this context, computational modelling and simulation of the cardiovascular system are crucial in furthering the understanding of CVDs, enhancing diagnostic capabilities, and developing patient-specific therapeutic interventions. Moreover, computer models can facilitate the development of novel medical devices, such as cardiac assist devices (CaAD) comprised of dielectric materials. In this thesis, the focus is on the development of two types of computational muscle models, specifically a cardiac muscle model based on biological tissue and an artificial muscle model based on dielectric materials. The work on the cardiac muscle model primarily focuses on the modelling of the orthotropic cardiac tissue structure. The developed tissue structure model (TSM) is based on a discontinuous Galerkin framework to accurately assess the transmural path and thickness in the myocardial wall. The framework enables more accurate modelling of the orthotropic tissue structure compared to established methods. Moreover, due to the modularity, the framework can be easily integrated into other TSMs. In this regard, transmural fibre and sheet angle rules are proposed for the left ventricle based on diffusion tensor magnetic resonance imaging (DT-MRI) data and demonstrate enhanced fidelity in representing the measurement data compared to existing rules. A study about the influence of different TSMs on important characteristics of cardiac function, based on an electromechanical model of the cardiac tissue, underscores the significant influence of the TSM. The artificial muscle model is inspired by the idea of an innovative concept for cardiac assist devices (CaAD) based on dielectric elastomers. However, there is a dearth of computational models that are able to simulate the complex electromechanical, dynamic, and viscoelastic behaviour of such a dielectric elastomer actuator-based CaAD. In this thesis, the computational model of the artificial muscle is based on an electromechanical shell formulation, including dynamics and viscoelasticity. The variational formulation of the dynamic, viscoelastic, and electromechanical shell is derived from the Lagrange-d’Alembert principle. A variational time integration ensures a good long-term energy behaviour. To demonstrate the potential of the model, numerical examples, including different geometries as well as deformation states, are presented. Overall, the proposed TSM, in conjunction with the improved transmural fibre and sheet angle rules, is a robust, efficient and accurate method by which to compute the orthotropic tissue structure for finite element models of the cardiac tissue. Furthermore, the significant influence of different TSMs on important characteristics of cardiac function is demonstrated. The electromechanical shell model proves to be a promising approach towards the development of patient-specific CaADs based on dielectric elastomer actuators (DEA).
  • Article
    Open Access
    Molekulardynamische Untersuchung von Zerkleinerungsprozessen in Kugelmühlen
    (Vogel Business Media, 2000) Buchholtz, Volkhard; Freund, Jan Alfred; Pöschel, Thorsten
    Molekulardynamische Untersuchungen eignen sich zur Simulation des Verhaltens makroskopischer Mengen granularen Materials unter bestimmten, technologisch relevanten Beanspruchungen. Wir untersuchen das Zerkleinerungsverhalten von Mahlgut in einer Kugelmühle, insbesondere im Hinblick auf die Effizienz als Funktion der Drehzahl und auf die räumliche Verteilung von Beanspruchungen. Die Untersuchung der Verteilung von Kraflketten (force chains) liefert eine Erklärung des Experiments von Rothkegel und Rolf [3].
  • Book part
    Open Access
    Die Kühlrate eines granularen Gases
    (Köster, 1996) Schwager, Thomas; Pöschel, Thorsten; Freund, Jan Alfred
  • Book part
    Open Access
    A measure of the information content of neural spike trains
    (DIRAC, 1997) Jiménez-Montaño, Miguel Angel; Pöschel, Thorsten; Rapp, Paul; Mizraji, Eduardo; Acerenza, Luis; Alvarez, Fabián; Pomi, Andrés
  • Conference object
    Open Access
    From “solid” to “fluid”: Time-dependent hydrodynamic analysis of dense granular flows
    (Kluwer Academic Press, 2001-09-30) Salueña, Clara; Pöschel, Thorsten; Esipov, Sergei E.; Aref, Hassan; Phillips, James W.
    When dealing with dense granular flows (not far above the ''fluidization point'' of the granular material), which cannot be regarded as granular gases, multiple unresolved questions arise. Many of them are related to the necessity of constructing the right framework to handle the dynamics of void occupation, which governs granular flow at high densities. This is a formidable task. However, hydrodynamic fields such as density, velocity, pressure and granular temperature, are easy to produce and study in numerical simulations of particles.
  • Conference object
    Open Access
    Entropy and Compressibility of Symbol Sequences
    (Complex Systems Institute, 1996-11-22) Ebeling, Werner; Pöschel, Thorsten; Neiman, Alexander
    The purpose of this paper is to investigate long-range correlations in symbol sequences using methods of statistical physics and nonlinear dynamics. Beside the principal interest in the analysis of correlations and fluctuations comprising many letters, our main aim is related here to the problem of sequence compression. In spite of the great progress in this field achieved in the work of Shannon, Fano, Huffman, Lempel, Ziv and others [1] many questions still remain open. In particular one must note that since the basic work by Lempel and Ziv the improvement of the standard compression algorithms is rather slow not exceeding a few percent per decade. One the other hand several experts expressed the idee that the long range correlations, which clearly exist in texts, computer programs etc. are not sufficiently taken into account by the standard algorithms [1]. Thus, our interest in compressibility is twofold: (i) We would like to explore how far compressibility is able to measure correlations. In particular we apply the standard algorithms to model sequences with known correlations such as, for instance, repeat sequences and symbolic sequences generated by maps. (ii) We aim to detect correlations which are not yet exploited in the standard compression algorithms and belong therefore to potential reservoirs for compression algorithms. First, the higher-order Shannon entropies are calculated. For repeat sequences analytic estimates are derived, which apply in some approximation to DNA sequences too. For symbolic strings obtained from special nonlinear maps and for several long texts a characteristic root law for the entropy scaling is detected. Then the compressibilities are estimated by using grammar representations and several standard computing algorithms. In particular we use the Lempel-Ziv compression algorithm. Further the mean square fluctuations of the composition with respect to the letter content and several characteristic scaling exponents are calculated. We show finally that all these measuresare able to detect long-range correlations. However, as demonstrated by shuffling experiments, different measuring methods operate on different length scales. The algorithms based on entropy or compressibility operate mainly on the word and sentence level. The characteristic scaling exponents reflect mainly the long-wave fluctuations of the composition which may comprise a few hundreds or thousands of letters.
  • Article
    Open Access
    Molecular Dynamics of Comminution in Ball Mills
    (Vogel Verlag, 2000) Buchholtz, Volkhard; Freund, Jan Alfred; Pöschel, Thorsten
    We Investigate autogenous fragmentation of dry granular material in rotating cylinders using two-dimensional molecular dynamics. By evaluation of spatial force distributions achieved numerically for various rotation velocities we argue that comminution occurs mainly due to the existence of force chains. A statistical analysis of these force chains explains the spatial distribution of comminution efficiency in ball mills as measured experimentally by Rothkegel [1] and Rolf [2].
  • Conference object
    Open Access
    Fluidization of granular material by vertical shaking
    (Kluwer Academic Press, 2001-09-30) Pöschel, Thorsten; Renard, Simon; Schwager, Thomas; Salueña, Clara; Aref, Hassan; Phillips, James W.
    For vertical shaken granular material we investigate the conditions for fluidization in the limits of a 1-d model. Whereas frequently the condition $A_0\omega_0^2/g\equif\Gamma>\Gamma_c$, ($A_0$: amplitude, $\omega_0$: frequency, $g$: gravity) is assumed, we derive a modified criterion which coincides with $\Gamma_c=1$ in the limit of small driving frequency $\omega_0$ only. Therefore, we conclude that the Froude number $\Gamma$ is not the proper criterion for the fluidization of a column of spheres and the more it is not suited to characterize a three dimensional granular material. The analytical results are supported by numerical and experimental data.
  • Conference object
    Open Access
    Rolling as a “continuing collision” for viscoelastic spheres
    (Kluwer Academic Press, 2001-09-30) Brilliantov, Nikolay; Pöschel, Thorsten; Aref, Hassan; Phillips, James W.
    We show that two basic mechanical processes, the collision of particles and rolling motion of a sphere on a plane, are intimately related. According to our recent findings, the restitution coefficient for colliding spherical particles $\varepsilon$, which characterizes the energy loss upon collision, is directly related to the rolling friction coefficient $\mu_\text{roll}$ for a viscous sphere on a hard plane. We quantify both coefficients in terms of material constants which allows to determine either of them provided the other is known. This relation between the coefficients may give rise to a novel experimental technique to determine alternatively the coefficient of restitution or the coefficient of rolling friction.
  • Book part
    Open Access
    Force Distribution and Comminution in Ball Mills
    (World Scientific, 1997-09-01) Buchholtz, Volkhard; Pöschel, Thorsten; Wolf, Dietrich E.; Grassberger, Peter
    The motion of granular material in a ball mill is investigated using molecular dynamics simulations in two dimensions. In agreement with experimental observations by Rothkegel [1] we find that local stresses – and hence the comminution efficiency – are maximal close to the bottom of the container. This effect will be explained using analysis of statistics of force chains in the material.
  • Book part
    Open Access
    Markovsche Analyse nasser Gemeinschaften
    (Logos, 2006) Freund, Jan Alfred; Pöschel, Thorsten; Wiltshire, Karen; Pöschel, Thorsten; Malchow, Horst; Schimansky-Geier, Lutz
    In ökologischen Systemen ist die saisonal variierende Präsenz von Arten sowohl durch die jahreszeitlich schwankenden Umgebungsbedingungen, als auch durch die Wechselwirkungen zwischen den Arten bedingt. Letztere erfahren ihre Ausprägung in Sukzession und Koexistenz. Wir wollen die zyklische Wiederkehr der Arten im Jahresgang wahrscheinlichkeitstheoretisch beschreiben, wobei der Wechsel von Jahr zu Jahr im Rahmen einer Markovschen Kette modelliert wird. Neben allgemeinen Ausführungen zu dieser modellhaften Art der Analyse werden wir als eine konkrete Anwendung die Zeitreihen (Zellzahlen) dreier prominenter Algenarten der südlichen Nordsee analysieren. Über den Aspekt einer quantitativen Beschreibung der Verhältnisse im marinen Habitat hinaus zielt eine derartige Untersuchung insbesondere auf einen Nachweis möglicher Veränderungen des ökologischen Systems der “nassen Gemeinschaft” vor dem Hintergrund einer bereits beobachteten Temperaturerhöhung des Meerwassers bei Helgoland.
  • Conference object
    Open Access
    Dynamic Entropies, Long-range Correlations and Fluctuations in Complex Linear Structures
    (World Scientific, 1995-12-01) Ebeling, Werner; Neiman, Alexander; Pöschel, Thorsten; Suzuki, Masuo; Kawashima, Naoki
    We investigate symbolic sequences and in particular information carriers as e.g. books and DNA{strings. First the higher order Shannon entropies are calculated, a characteristic root law is detected. Then the algorithmic entropy is estimated by using Lempel-Ziv compression algorithms. In the third section the correlation function for distant letters, the low frequency Fourier spectrum and the characteristic scaling exponents are calculated. We show that all these measures are able to detect long{range correlations. However, as demonstrated by shuffling experiments, different measures operate on different length scales. The longest correlations found in our analysis comprise a few hundreds or thousands of letters and may be understood as long-wave fluctuations of the composition.
  • Book part
    Open Access
    Kann die Entropie von Sequenzen vermittelst der Kompressibilität gemessen werden?
    (Köster, 1996) Pöschel, Thorsten; Freund, Jan Alfred
  • Book part
    Open Access
    Efficient numerical simulation of granular matter using the Bottom-To-Top Reconstruction method
    (Shaker, 2006-11-08) Schwager, Thomas; Pöschel, Thorsten; Walzel, Peter; Grochowski, Rafał; Krülle, Christof A.; Linz, Stefan J.
    The numerical simulation of granular systems of even moderate size is a challenging computational problem. In most investigations, either Molecular Dynamics or Event-driven Molecular Dynamics is applied. Here we show that in certain cases, mainly (but not exclusively) for static granular packings, the Bottom-to-top Reconstruction method allows for the efficient simulation of very large systems. We apply the method to heap formation, granular flow in a rotating cylinder and to structure formation in nano-powders. We also present an efficient implementation of the algorithm in C++, including a benchmark.
  • Conference object
    Open Access
    Hydrodynamic fluctuations and averaging problems in dense granular flows
    (Balkema, 1997) Saluneña, Clara; Esipov, Sergei E.; Pöschel, Thorsten; Behringer, Robert P.; Jenkins, James T.
    We analyze the properties of dense granular systems by assuming a hydrodynamical description based on conservation laws for the particle number density and linear momentum. We combine analytical methods and experimental and numerical results obtained by ensemble-averaging of data on creep during compaction and molecular dynamics simulations of convective flow.
  • Book part
    Open Access
    Contact of Viscoelastic Spheres
    (World Scientific, 1997-09-01) Schwager, Thomas; Pöschel, Thorsten; Wolf, Dietrich E.; Grassberger, Peter
    In a recent paper an implicit equation for contacting viscoelastic spheres was derived [1]. Integrating this equation it can be shown that the coefficient of normal restitution $\varepsilon$ depends on the impact velocity, $g$, as as $1 − \varepsilon \propto g^{1/5}$.
  • Doctoral thesis
    Open Access
    Automating the Design of Multigrid Methods with Evolutionary Program Synthesis
    FAU Studien aus der Informatik : 19, (FAU University Press, 2024) Schmitt, Jonas; Köstler, Harald
    Many of the most fundamental laws of nature can be formulated as partial differential equations (PDEs). However, since the general solution of many PDEs is unknown, the efficient approximate solution of these equations is one of humanity's greatest challenges. While multigrid represents one of the most effective methods for solving PDEs numerically, in many cases, the design of an efficient or at least working multigrid solver is an open problem. This thesis demonstrates that grammar-guided genetic programming, an evolutionary program synthesis technique, can discover multigrid methods of unprecedented structure that achieve a high degree of efficiency and generalization. For this purpose, we develop a novel context-free grammar that enables the automated generation of multigrid methods in a symbolically-manipulable formal language, based on which we can apply the same multigrid-based solver to problems of different sizes without having to adapt its internal structure. Treating the automated design of an efficient multigrid method as a program synthesis task allows us to find novel sequences of multigrid operations, including the combination of different smoothing and coarse-grid correction steps on each level of the discretization hierarchy.
  • Book part
    Open Access
    Mechanik körniger Medien
    (Wiley-VCH, 2003) Gudehus, Gerd; Herrmann, Hans; Hutter, Kolumban; Pöschel, Thorsten; Schwedes, Jörg