Reconstrução de imagem médica de mamografia por emissão de positrões (PEM) com GPU

Dissertação para obtenção do Grau de Mestre em Engenharia Biomédica

Heterogeneous computing with an algorithmic skeleton framework

The Graphics Processing Unit (GPU) is present in almost every modern day personal computer. Despite its specific purpose design, they have been increasingly used for general computations with very good results. Hence, there is a growing effort from the community to seamlessly integrate this kind of devices in everyday computing. However, to fully exploit the potential of a system comprising GPUs and CPUs, these devices should be presented to the programmer as a single platform. The efficient combination of the power of CPU and GPU devices is highly dependent on each device’s characteristics, resulting in platform specific applications that cannot be ported to different systems. Also, the most efficient work balance among devices is highly dependable on the computations to be performed and respective data sizes. In this work, we propose a solution for heterogeneous environments based on the abstraction level provided by algorithmic skeletons. Our goal is to take full advantage of the power of all CPU and GPU devices present in a system, without the need for different kernel implementations nor explicit work-distribution.To that end, we extended Marrow, an algorithmic skeleton framework for multi-GPUs, to support CPU computations and efficiently balance the work-load between devices. Our approach is based on an offline training execution that identifies the ideal work balance and platform configurations for a given application and input data size. The evaluation of this work shows that the combination of CPU and GPU devices can significantly boost the performance of our benchmarks in the tested environments, when compared to GPU-only executions.

Técnicas de paralelização em GPGPU aplicadas em algoritmo para remoção de ruído multiplicativo

Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)

Implementação do algoritmo de treinamento do classificador Floresta de Caminhos Ótimos em GPU

Pós-graduação em Ciência da Computação - IBILCE

Implementação do algoritmo de treinamento do classificador floresta de caminhos ótimos em GPU

Técnicas de reconhecimento de padrões tem como principal objetivo classificar um conjunto de amostras, sendo o processo de aprendizado a fase de maior consumo de tempo. O problema pode piorar em ferramentas de classificação interativas, o que pode ser inaceitável para grandes bases de dados. Um exemplo de classificador é o baseado em Floresta de Caminhos Ótimos [8] - OPF. Dado que muitos trabalhos tem sido orientados à implementação de algoritmos de reconhecimento de padrões em ambiente General Purpose Graphics Processing Unit - GPGPU, o presente estudo objetivou a implementação da etapa de treinamento do classificador Floresta de Caminhos Ótimos em CUDA, visando aumentar a sua eficiência. A otimização do classificador em CUDA demonstrou uma fase de treinamento mais rápida que a versão original.

Paralelización de un algoritmo de ray tracing para arrays de mililentes

Máster Universitario en Sistemas Inteligentes y Aplicaciones Numéricas en Ingeniería (SIANI)

Increasing Realism and Supporting Content Planning for Dynamic Scenes in a Mixed Reality System incorporating a Time-of-Flight Camera

For broadcasting purposes MIXED REALITY, the combination of real and virtual scene content, has become ubiquitous nowadays. Mixed Reality recording still requires expensive studio setups and is often limited to simple color keying. We present a system for Mixed Reality applications which uses depth keying and provides threedimensional mixing of real and artificial content. It features enhanced realism through automatic shadow computation which we consider a core issue to obtain realism and a convincing visual perception, besides the correct alignment of the two modalities and correct occlusion handling. Furthermore we present a possibility to support placement of virtual content in the scene. Core feature of our system is the incorporation of a TIME-OF-FLIGHT (TOF)-camera device. This device delivers real-time depth images of the environment at a reasonable resolution and quality. This camera is used to build a static environment model and it also allows correct handling of mutual occlusions between real and virtual content, shadow computation and enhanced content planning. The presented system is inexpensive, compact, mobile, flexible and provides convenient calibration procedures. Chroma-keying is replaced by depth-keying which is efficiently performed on the GRAPHICS PROCESSING UNIT (GPU) by the usage of an environment model and the current ToF-camera image. Automatic extraction and tracking of dynamic scene content is herewith performed and this information is used for planning and alignment of virtual content. An additional sustainable feature is that depth maps of the mixed content are available in real-time, which makes the approach suitable for future 3DTV productions. The presented paper gives an overview of the whole system approach including camera calibration, environment model generation, real-time keying and mixing of virtual and real content, shadowing for virtual content and dynamic object tracking for content planning.

Non-rigid Free-Form 2D-3D Registration Using Statistical Deformation Model

This paper presents a non-rigid free-from 2D-3D registration approach using statistical deformation model (SDM). In our approach the SDM is first constructed from a set of training data using a non-rigid registration algorithm based on b-spline free-form deformation to encode a priori information about the underlying anatomy. A novel intensity-based non-rigid 2D-3D registration algorithm is then presented to iteratively fit the 3D b-spline-based SDM to the 2D X-ray images of an unseen subject, which requires a computationally expensive inversion of the instantiated deformation in each iteration. In this paper, we propose to solve this challenge with a fast B-spline pseudo-inversion algorithm that is implemented on graphics processing unit (GPU). Experiments conducted on C-arm and X-ray images of cadaveric femurs demonstrate the efficacy of the present approach.

Moving object detection for real-time augmented reality applications in a GPGPU

The last generation of consumer electronic devices is endowed with Augmented Reality (AR) tools. These tools require moving object detection strategies, which should be fast and efficient, to carry out higher level object analysis tasks. We propose a lightweight spatio-temporal-based non-parametric background-foreground modeling strategy in a General Purpose Graphics Processing Unit (GPGPU), which provides real-time high-quality results in a great variety of scenarios and is suitable for AR applications.

GPGPU implementation of growing neural gas: application to 3D scene reconstruction

Self-organising neural models have the ability to provide a good representation of the input space. In particular the Growing Neural Gas (GNG) is a suitable model because of its flexibility, rapid adaptation and excellent quality of representation. However, this type of learning is time-consuming, especially for high-dimensional input data. Since real applications often work under time constraints, it is necessary to adapt the learning process in order to complete it in a predefined time. This paper proposes a Graphics Processing Unit (GPU) parallel implementation of the GNG with Compute Unified Device Architecture (CUDA). In contrast to existing algorithms, the proposed GPU implementation allows the acceleration of the learning process keeping a good quality of representation. Comparative experiments using iterative, parallel and hybrid implementations are carried out to demonstrate the effectiveness of CUDA implementation. The results show that GNG learning with the proposed implementation achieves a speed-up of 6× compared with the single-threaded CPU implementation. GPU implementation has also been applied to a real application with time constraints: acceleration of 3D scene reconstruction for egomotion, in order to validate the proposal.

High performance numerical modeling of ultra-short laser pulse propagation based on multithreaded parallel hardware

The focus of this study is development of parallelised version of severely sequential and iterative numerical algorithms based on multi-threaded parallel platform such as a graphics processing unit. This requires design and development of a platform-specific numerical solution that can benefit from the parallel capabilities of the chosen platform. Graphics processing unit was chosen as a parallel platform for design and development of a numerical solution for a specific physical model in non-linear optics. This problem appears in describing ultra-short pulse propagation in bulk transparent media that has recently been subject to several theoretical and numerical studies. The mathematical model describing this phenomenon is a challenging and complex problem and its numerical modeling limited on current modern workstations. Numerical modeling of this problem requires a parallelisation of an essentially serial algorithms and elimination of numerical bottlenecks. The main challenge to overcome is parallelisation of the globally non-local mathematical model. This thesis presents a numerical solution for elimination of numerical bottleneck associated with the non-local nature of the mathematical model. The accuracy and performance of the parallel code is identified by back-to-back testing with a similar serial version.

Towards an algorithmic skeleton framework for programming the Intel R Xeon PhiTM processor

The Intel R Xeon PhiTM is the first processor based on Intel’s MIC (Many Integrated Cores) architecture. It is a co-processor specially tailored for data-parallel computations, whose basic architectural design is similar to the ones of GPUs (Graphics Processing Units), leveraging the use of many integrated low computational cores to perform parallel computations. The main novelty of the MIC architecture, relatively to GPUs, is its compatibility with the Intel x86 architecture. This enables the use of many of the tools commonly available for the parallel programming of x86-based architectures, which may lead to a smaller learning curve. However, programming the Xeon Phi still entails aspects intrinsic to accelerator-based computing, in general, and to the MIC architecture, in particular. In this thesis we advocate the use of algorithmic skeletons for programming the Xeon Phi. Algorithmic skeletons abstract the complexity inherent to parallel programming, hiding details such as resource management, parallel decomposition, inter-execution flow communication, thus removing these concerns from the programmer’s mind. In this context, the goal of the thesis is to lay the foundations for the development of a simple but powerful and efficient skeleton framework for the programming of the Xeon Phi processor. For this purpose we build upon Marrow, an existing framework for the orchestration of OpenCLTM computations in multi-GPU and CPU environments. We extend Marrow to execute both OpenCL and C++ parallel computations on the Xeon Phi. We evaluate the newly developed framework, several well-known benchmarks, like Saxpy and N-Body, will be used to compare, not only its performance to the existing framework when executing on the co-processor, but also to assess the performance on the Xeon Phi versus a multi-GPU environment.

Algorithms hardware implementation for ultrasonic data processing in SHM system

Nowadays, devices that monitor the health of structures consume a lot of power and need a lot of time to acquire, process, and send the information about the structure to the main processing unit. To decrease this time, fast electronic devices are starting to be used to accelerate this processing. In this paper some hardware algorithms implemented in an electronic logic programming device are described. The goal of this implementation is accelerate the process and diminish the information that has to be send. By reaching this goal, the time the processor needs for treating all the information is reduced and so the power consumption is reduced too.

Aceleração GPU da animação de superfícies deformáveis

Dissertação de Mestrado em Engenharia Informática

Scalable and Efficient Configuration of TimeDivision Multiplexed Resources

Consumer-electronics systems are becoming increasingly complex as the number of integrated applications is growing. Some of these applications have real-time requirements, while other non-real-time applications only require good average performance. For cost-efficient design, contemporary platforms feature an increasing number of cores that share resources, such as memories and interconnects. However, resource sharing causes contention that must be resolved by a resource arbiter, such as Time-Division Multiplexing. A key challenge is to configure this arbiter to satisfy the bandwidth and latency requirements of the real-time applications, while maximizing the slack capacity to improve performance of their non-real-time counterparts. As this configuration problem is NP-hard, a sophisticated automated configuration method is required to avoid negatively impacting design time. The main contributions of this article are: 1) An optimal approach that takes an existing integer linear programming (ILP) model addressing the problem and wraps it in a branch-and-price framework to improve scalability. 2) A faster heuristic algorithm that typically provides near-optimal solutions. 3) An experimental evaluation that quantitatively compares the branch-and-price approach to the previously formulated ILP model and the proposed heuristic. 4) A case study of an HD video and graphics processing system that demonstrates the practical applicability of the approach.