Large-scale Network Analysis on Distributed Architectures

Questa dissertazione esamina le sfide e i limiti che gli algoritmi di analisi di grafi incontrano in architetture distribuite costituite da personal computer. In particolare, analizza il comportamento dell'algoritmo del PageRank così come implementato in una popolare libreria C++ di analisi di grafi distribuiti, la Parallel Boost Graph Library (Parallel BGL). I risultati qui presentati mostrano che il modello di programmazione parallela Bulk Synchronous Parallel è inadatto all'implementazione efficiente del PageRank su cluster costituiti da personal computer. L'implementazione analizzata ha infatti evidenziato una scalabilità negativa, il tempo di esecuzione dell'algoritmo aumenta linearmente in funzione del numero di processori. Questi risultati sono stati ottenuti lanciando l'algoritmo del PageRank della Parallel BGL su un cluster di 43 PC dual-core con 2GB di RAM l'uno, usando diversi grafi scelti in modo da facilitare l'identificazione delle variabili che influenzano la scalabilità. Grafi rappresentanti modelli diversi hanno dato risultati differenti, mostrando che c'è una relazione tra il coefficiente di clustering e l'inclinazione della retta che rappresenta il tempo in funzione del numero di processori. Ad esempio, i grafi Erdős–Rényi, aventi un basso coefficiente di clustering, hanno rappresentato il caso peggiore nei test del PageRank, mentre i grafi Small-World, aventi un alto coefficiente di clustering, hanno rappresentato il caso migliore. Anche le dimensioni del grafo hanno mostrato un'influenza sul tempo di esecuzione particolarmente interessante. Infatti, si è mostrato che la relazione tra il numero di nodi e il numero di archi determina il tempo totale.

A parallel community detection algorithm

Complex networks analysis is a very popular topic in computer science. Unfortunately this networks, extracted from different contexts, are usually very large and the analysis may be very complicated: computation of metrics on these structures could be very complex. Among all metrics we analyse the extraction of subnetworks called communities: they are groups of nodes that probably play the same role within the whole structure. Communities extraction is an interesting operation in many different fields (biology, economics,...). In this work we present a parallel community detection algorithm that can operate on networks with huge number of nodes and edges. After an introduction to graph theory and high performance computing, we will explain our design strategies and our implementation. Then, we will show some performance evaluation made on a distributed memory architectures i.e. the supercomputer IBM-BlueGene/Q "Fermi" at the CINECA supercomputing center, Italy, and we will comment our results.

Numerical simulation of interdigitated back contact hetero-junction solar cells

The goal of this thesis is the application of an opto-electronic numerical simulation to heterojunction silicon solar cells featuring an all back contact architecture (Interdigitated Back Contact Hetero-Junction IBC-HJ). The studied structure exhibits both metal contacts, emitter and base, at the back surface of the cell with the objective to reduce the optical losses due to the shadowing by front contact of conventional photovoltaic devices. Overall, IBC-HJ are promising low-cost alternatives to monocrystalline wafer-based solar cells featuring front and back contact schemes, in fact, for IBC-HJ the high concentration doping diffusions are replaced by low-temperature deposition processes of thin amorphous silicon layers. Furthermore, another advantage of IBC solar cells with reference to conventional architectures is the possibility to enable a low-cost assembling of photovoltaic modules, being all contacts on the same side. A preliminary extensive literature survey has been helpful to highlight the specific critical aspects of IBC-HJ solar cells as well as the state-of-the-art of their modeling, processing and performance of practical devices. In order to perform the analysis of IBC-HJ devices, a two-dimensional (2-D) numerical simulation flow has been set up. A commercial device simulator based on finite-difference method to solve numerically the whole set of equations governing the electrical transport in semiconductor materials (Sentuarus Device by Synopsys) has been adopted. The first activity carried out during this work has been the definition of a 2-D geometry corresponding to the simulation domain and the specification of the electrical and optical properties of materials. In order to calculate the main figures of merit of the investigated solar cells, the spatially resolved photon absorption rate map has been calculated by means of an optical simulator. Optical simulations have been performed by using two different methods depending upon the geometrical features of the front interface of the solar cell: the transfer matrix method (TMM) and the raytracing (RT). The first method allows to model light prop-agation by plane waves within one-dimensional spatial domains under the assumption of devices exhibiting stacks of parallel layers with planar interfaces. In addition, TMM is suitable for the simulation of thin multi-layer anti reflection coating layers for the reduction of the amount of reflected light at the front interface. Raytracing is required for three-dimensional optical simulations of upright pyramidal textured surfaces which are widely adopted to significantly reduce the reflection at the front surface. The optical generation profiles are interpolated onto the electrical grid adopted by the device simulator which solves the carriers transport equations coupled with Poisson and continuity equations in a self-consistent way. The main figures of merit are calculated by means of a postprocessing of the output data from device simulation. After the validation of the simulation methodology by means of comparison of the simulation result with literature data, the ultimate efficiency of the IBC-HJ architecture has been calculated. By accounting for all optical losses, IBC-HJ solar cells result in a theoretical maximum efficiency above 23.5% (without texturing at front interface) higher than that of both standard homojunction crystalline silicon (Homogeneous Emitter HE) and front contact heterojuction (Heterojunction with Intrinsic Thin layer HIT) solar cells. However it is clear that the criticalities of this structure are mainly due to the defects density and to the poor carriers transport mobility in the amorphous silicon layers. Lastly, the influence of the most critical geometrical and physical parameters on the main figures of merit have been investigated by applying the numerical simulation tool set-up during the first part of the present thesis. Simulations have highlighted that carrier mobility and defects level in amorphous silicon may lead to a potentially significant reduction of the conversion efficiency.

Parallelization of the algorithm WHAM with NVIDIA CUDA.

The aim of my thesis is to parallelize the Weighting Histogram Analysis Method (WHAM), which is a popular algorithm used to calculate the Free Energy of a molucular system in Molecular Dynamics simulations. WHAM works in post processing in cooperation with another algorithm called Umbrella Sampling. Umbrella Sampling has the purpose to add a biasing in the potential energy of the system in order to force the system to sample a specific region in the configurational space. Several N independent simulations are performed in order to sample all the region of interest. Subsequently, the WHAM algorithm is used to estimate the original system energy starting from the N atomic trajectories. The parallelization of WHAM has been performed through CUDA, a language that allows to work in GPUs of NVIDIA graphic cards, which have a parallel achitecture. The parallel implementation may sensibly speed up the WHAM execution compared to previous serial CPU imlementations. However, the WHAM CPU code presents some temporal criticalities to very high numbers of interactions. The algorithm has been written in C++ and executed in UNIX systems provided with NVIDIA graphic cards. The results were satisfying obtaining an increase of performances when the model was executed on graphics cards with compute capability greater. Nonetheless, the GPUs used to test the algorithm is quite old and not designated for scientific calculations. It is likely that a further performance increase will be obtained if the algorithm would be executed in clusters of GPU at high level of computational efficiency. The thesis is organized in the following way: I will first describe the mathematical formulation of Umbrella Sampling and WHAM algorithm with their apllications in the study of ionic channels and in Molecular Docking (Chapter 1); then, I will present the CUDA architectures used to implement the model (Chapter 2); and finally, the results obtained on model systems will be presented (Chapter 3).