Unsupervised Markovian Segmentation on Graphics Hardware

Graphics processing units (GPUs) are providing increasingly higher performance with programmable internal processors, namely vertex processors (VPs) and fragment processors (FPs). Such newly added capabilities motivate us to perform general-purpose computation on GPUs (GPGPU) beyond graphics applications. Although VPs and FPs are connected in a pipeline, many GPGPU implementations utilize only FPs as a computational engine in the GPU. Therefore, such implementations may result in lower performance due to highly loaded FPs (as compared to VPs) being a performance bottleneck in the pipeline execution. The objective of our work is to improve the performance of GPGPU programs by eliminating this bottleneck. To achieve this, we present a code motion technique that is capable of reducing the FP workload by moving assembly instructions appropriately from the FP program to the VP program. We also present the definition of such movable instructions that do not change the I/O specification between the CPU and the GPU. The experimental results show that (1) our technique improves the performance of a Gaussian filter program with reducing execution time by approximately 40% and (2) it successfully reduces the FP workload in 10 out of 18 GPGPU programs.

Graphic processing units (GPUs) emerged recently as an exciting new hardware environment for a truly parallel implementation and execution of Nature and Bio-inspired Algorithms with excellent price-to-power ratio. In contrast to common multicore CPUs that contain up to tens of independent cores, the GPUs represent a massively parallel single-instruction multiple-data devices that can nowadays reach peak performance of hundreds and thousands of giga floating-point operations per second. Nature and Bio-inspired Algorithms implement parallel optimization strategies in which a single candidate solution, a group of candidate solutions (population), or multiple populations seek for optimal solution or set of solutions of given problem. Genetic algorithms (GA) constitute a family of traditional and very well-known nature-inspired populational meta-heuristic algorithms that have proved its usefulness on a plethora of tasks through the years. Differential evolution (DE) is another efficient populational meta-heuristic algorithm for real-parameter optimization. Particle swarm optimization (PSO) can be seen as nature-inspired multiagent method in which the interaction of simple independent agents yields intelligent collective behavior. Simulated annealing (SA) is global optimization algorithm which combines statistical mechanics and combinatorial optimization with inspiration in metallurgy. This survey provides a brief overview of the latest state-of-the-art research on the design, implementation, and applications of parallel GA, DE, PSO, and SA-based methods on the GPUs.

In this work we propose a highly optimized version of a simulated annealing (SA) algorithm adapted to the more recently developed graphic processor units (GPUs). The programming has been carried out with compute unified device architecture (CUDA) toolkit, specially designed for Nvidia GPUs. For this purpose, efficient versions of SA have been first analyzed and adapted to GPUs. Thus, an appropriate sequential SA algorithm has been developed as starting point. Next, a straightforward asynchronous parallel version has been implemented and then a specific and more efficient synchronous version has been developed. A wide appropriate benchmark to illustrate the performance properties of the implementation has been considered. Among all tests, a classical sample problem provided by the minimization of the normalized Schwefel function has been selected to compare the behavior of the sequential, asynchronous and synchronous versions, the last one being more advantageous in terms of balance between convergence, accuracy and computational cost. Also the implementation of a hybrid method combining SA with a local minimizer method has been developed. Note that the generic feature of the SA algorithm allows its application in a wide set of real problems arising in a large variety of fields, such as biology, physics, engineering, finance and industrial processes.

SummaryAn OpenACC directive‐based graphics processing unit (GPU) parallel scheme is presented for solving the compressible Navier–Stokes equations on 3D hybrid unstructured grids with a third‐order reconstructed discontinuous Galerkin method. The developed scheme requires the minimum code intrusion and algorithm alteration for upgrading a legacy solver with the GPU computing capability at very little extra effort in programming, which leads to a unified and portable code development strategy. A face coloring algorithm is adopted to eliminate the memory contention because of the threading of internal and boundary face integrals. A number of flow problems are presented to verify the implementation of the developed scheme. Timing measurements were obtained by running the resulting GPU code on one Nvidia Tesla K20c GPU card (Nvidia Corporation, Santa Clara, CA, USA) and compared with those obtained by running the equivalent Message Passing Interface (MPI) parallel CPU code on a compute node (consisting of two AMD Opteron 6128 eight‐core CPUs (Advanced Micro Devices, Inc., Sunnyvale, CA, USA)). Speedup factors of up to 24× and 1.6× for the GPU code were achieved with respect to one and 16 CPU cores, respectively. The numerical results indicate that this OpenACC‐based parallel scheme is an effective and extensible approach to port unstructured high‐order CFD solvers to GPU computing. Copyright © 2015 John Wiley & Sons, Ltd.

Solid modeling of polyhedral objects by Layered Depth-Normal Images on the GPU

BackgroundSearching a database of protein structures for matches to a query structure, or occurrences of a structural motif, is an important task in structural biology and bioinformatics. While there are many existing methods for structural similarity searching, faster and more accurate approaches are still required, and few current methods are capable of substructure (motif) searching.ResultsWe developed an improved heuristic for tableau-based protein structure and substructure searching using simulated annealing, that is as fast or faster and comparable in accuracy, with some widely used existing methods. Furthermore, we created a parallel implementation on a modern graphics processing unit (GPU).ConclusionsThe GPU implementation achieves up to 34 times speedup over the CPU implementation of tableau-based structure search with simulated annealing, making it one of the fastest available methods. To the best of our knowledge, this is the first application of a GPU to the protein structural search problem.

This study presents an asynchronous parallel strategy coordinating central processing unit (CPU) and graphic processing unit (GPU) to accelerate neighborhood operation (NO). Specifically, we propose a data partitioning method called multi-anchor task queuing and a task scheduling method called bi-direction task scheduling, which can support CPU and GPU to find the responsible data blocks rapidly and concurrently handle their tasks via a bi-direction merge. Moreover, we optimize the organization of threads distributed among the CPU and GPU. Experimental results show that when a 1.7 GB raster dataset is processed, the speedup ratio achieved by the proposed parallel algorithm reaches 29.63, which is 19% and 18% higher than those of the GPU and standard asynchronous parallel algorithm, respectively. Additionally, the load balance index is below 0.085, which is significantly better than the value achieved by a conventional algorithm. Thus, the strategy achieves a higher speedup ratio and more adaptable load balance, thereby accelerating the NO more efficiently. Further, the impacts of the data volume, computational intensity, organization mode of the GPU threads, and granularity of the GPU stream on the parallel efficiency are evaluated and discussed. We also test the efficiency of four other common NOs with our strategy.

Integrating batched sparse iterative solvers for the collision operator in fusion plasma simulations on GPUs

Segmentation is a low-level processing aimed at the partition of an image in disjoint regions, each one homogeneous with respect to some properties like intensity, texture, shape, etc. Such a task is needed in many high-level processing and applications in such diverse fields as remote-sensing, medical imaging, video coding, and industrial automation, just to name a few. In this work of thesis, we first deal with the problem of color based image segmentation, that is based on some form of homogeneity in the color/spectral properties of the image, and then move to the more complex task of texture based segmentation, where the aim is to recognize complex structures in the image which are typically non homogeneous in terms of spectral properties. Concerning color based segmentation, we focus on a specific class of Markov Random Field (MRF) probability models, namely the Tree-Structured Markov Random Fields (TS-MRF), recently proposed to provide a hierarchical multiscale description of images. Based on such a model, the unsupervised image segmentation is carried out by means of a sequence of nested class splits, where each class is modeled as a local binary MRF. As a first contribution, we first extend the model to generic tree structures removing the unnecessary binary constraint, and then devise a new TS-MRF unsupervised segmentation technique which improves upon the original algorithm by selecting a better tree structure and eliminating spurious classes. Such results are obtained by using the Mean-Shift procedure to estimate the number of pdf modes at each node, and to obtain a more reliable initial clustering for subsequent MRF optimization. To this end, we devise a new reliable and fast clustering algorithm based on the Mean-Shift technique, which makes use of a variable-bandwidth strategy based on the k-Nearest Neighbors (k-NN) technique, and is implemented with a speed-up strategy that cuts significantly the computational complexity. Experimental results on synthetic data and real life images from the remote sensing domain prove the potential of the proposed method. Turning to texture based segmentation, we present a novel multiscale texture model, and a related algorithm for the unsupervised segmentation of color images. Elementary textures are characterized by their spatial interactions with neighboring regions along selected directions. Such interactions are modeled in turn by means of a set of Markov chains, one for each direction, whose parameters are collected in a feature vector that synthetically describes the texture. Based on the feature vectors, the texture are then recursively merged, giving rise to larger and more complex textures, which appear at different scales of observation: accordingly, the model is named Hierarchical Multiple Markov Chain (H-MMC). The Texture Fragmentation and Reconstruction (TFR) algorithm, addresses the unsupervised segmentation problem based on the H-MMC model. The ``fragmentation'' step allows one to find the elementary textures of the model, while the ``reconstruction'' step defines the hierarchical image segmentation based on a probabilistic measure (texture score) which takes into account both region scale and inter-region interactions. The performance of the proposed method was assessed through the Prague segmentation benchmark, based on mosaics of real natural textures, and also tested on real-world natural images from the Berkeley dataset. Finally, an application of the TFR algorithm for the unsupervised segmentation of multiresolution remotely sensed images is proposed, based on the significant presence of textural content in last generation high-resolution satellite images. The obtained results further assess the potential of the proposed method for such real life applications.

Purpose:The calculation time is a trade‐off for improving the accuracy of convolution dose calculation with fine calculation spacing of the KERMA kernel. We investigated to accelerate the convolution calculation using an ideal kernel on the Graphic Processing Units (GPU).Methods:The calculation was performed on the AMD graphics hardware of Dual FirePro D700 and our algorithm was implemented using the Aparapi that convert Java bytecode to OpenCL. The process of dose calculation was separated with the TERMA and KERMA steps. The dose deposited at the coordinate (x, y, z) was determined in the process. In the dose calculation running on the central processing unit (CPU) of Intel Xeon E5, the calculation loops were performed for all calculation points. On the GPU computation, all of the calculation processes for the points were sent to the GPU and the multi‐thread computation was done. In this study, the dose calculation was performed in a water equivalent homogeneous phantom with 1503 voxels (2 mm calculation grid) and the calculation speed on the GPU to that on the CPU and the accuracy of PDD were compared.Results:The calculation time for the GPU and the CPU were 3.3 sec and 4.4 hour, respectively. The calculation speed for the GPU was 4800 times faster than that for the CPU. The PDD curve for the GPU was perfectly matched to that for the CPU.Conclusion:The convolution calculation with the ideal kernel on the GPU was clinically acceptable for time and may be more accurate in an inhomogeneous region. Intensity modulated arc therapy needs dose calculations for different gantry angles at many control points. Thus, it would be more practical that the kernel uses a coarse spacing technique if the calculation is faster while keeping the similar accuracy to a current treatment planning system.

Graphics Processing Units (GPU) are becoming a widespread tool for general-purpose scientific computing, and are attracting interest for future onboard satellite image processing payloads due to their ability to perform massively parallel computations. This paper describes the GPU implementation of an algorithm for onboard lossy hyperspectral image compression, and proposes an architecture that allows to accelerate the compression task by parallelizing it on the GPU. The selected algorithm was amenable to parallel computation owing to its block-based operation, and has been optimized here to facilitate GPU implementation incurring a negligible overhead with respect to the original single-threaded version. In particular, a parallelization strategy has been designed for both the compressor and the corresponding decompressor, which are implemented on a GPU using Nvidia's CUDA parallel architecture. Experimental results on several hyperspectral images with different spatial and spectral dimensions are presented, showing significant speed-ups with respect to a single-threaded CPU implementation. These results highlight the significant benefits of GPUs for onboard image processing, and particularly image compression, demonstrating the potential of GPUs as a future hardware platform for very high data rate instruments.

This study aims to employ both central processing units (CPUs) and graphics processing units (GPUs) to collaboratively generate contour lines from a large-scale digital elevation model (DEM). The performance was improved with regard to three aspects. First, the original DEM data were decomposed and assigned according to the GPU’s limited memory so that large-scale data could be correctly addressed. Second, different types of computational tasks between the CPUs and GPUs were dynamically scheduled to ensure that both accelerators cooperate for performance improvement. Third, parallel processing on GPUs and CPUs was separately optimized for more efficient acceleration. Experimental results indicated that applying the parallel algorithm to data with a volume of 37.81 GB and area of 5,975,625.16 km2 reduced the total execution time from 332.84 min to 8.29 min for an optimal speedup of 40.15. In addition, we investigated the effects of the computational intensity, decomposition granularity, and task scheduling on parallel efficiency and performance improvement. We also discussed its degree of effectiveness, broader application, and the future direction of research in the field of geographic information systems.

A new Graphics Processing Unit (GPU) parallelization strategy is proposed to accelerate sparse finite element computation for three dimensional electromagnetic analysis. The parallelization strategy is employed based on a new compression format called sliced ELL Four (sliced ELL-F). The sliced ELL-F format-based parallelization strategy is designed for hastening many addition, dot product, and Sparse Matrix Vector Product (SMVP) operations in the Conjugate Gradient Norm (CGN) calculation of finite element equations. The new implementation of SMVP on GPUs is evaluated. The proposed strategy executed on a GPU can efficiently solve sparse finite element equations, especially when the equations are huge sparse (size of most rows in a coefficient matrix is less than 8). Numerical results show the sliced ELL-F format-based parallelization strategy can reach significant speedups compared to Compressed Sparse Row (CSR) format.

Multi-phase SPH modelling of violent hydrodynamics on GPUs

In this paper, a recent unsupervised segmentation algorithm based on computation of 2-dimensional (2D) local center of mass (LCM) technique is used to segment X-ray images of Knee Osteoarthritis (KOA) of various stages. Clinically, KOA is classified into four grades ranging from 1 to 4 based on the degradation of ligament in between the two bones fibula and femur and causes suffering with impaired movement. Identifying this space between bones through anterior view of knee X-ray image can indicate the type of KOA and helps in selection of suitable treatment process. Several segmentations methods in scrutinizing the 2D Knee X-ray image for KOA is in practice are supervised segmentation methods. Unsupervised segmentation techniques like watershed, K-means etc., are not sufficient to produce a precise segmentation results on its own and also segmented image takes more memory space. In this research, a new unsupervised segmentation algorithm based on computation of LCM technique is used to segment 2D knee X-ray images to screen the KOA stages. The calculated center of masses (CM) is related to the spatial region of pixels and forms the clusters; hence the segmentation is done based on the information from the whole image rather than only on the neighboring pixels. LCM method of segmentation is tested on 80 clinically categorized X-ray images (20 from each category) of KOA. The results are compared with two popular segmentation methods such as Watershed and Gaussian Mixed Model-Hidden Markov Random Field (GMM-HMRF) segmentation methods, respectively. Evaluated Dice score gives higher value of 0.59 ± 0.02 for the segmented results obtained from LCM computation method. Memory space occupied by LCM based segmented images is around 250 ± 30 kilo bytes (KB) which is less compared to other two watershed and GMM-HMRF methods. Overall, it is observed that LCM method could able to produce a crystal-clear segmented region boundaries and less memory space when compared to other two methods. The CMs calculated also depends on the intensity value of grey scale pixels and this method would be more appropriate for bone mineral density (BMD) analysis.