Programmable Graphics Processing Units Research Articles

GPU-Accelerated Dynamic Wavelet Thresholding Algorithm for X-Ray CT Metal Artifact Reduction

The computed tomography (CT) imaging technique has been used to diagnose widespread disease for many years; however, severe streaking artifacts have been observed to appear in the reconstructed images of examinate contains metal objects. In this paper, we propose a dynamic wavelet thresholding metal artifact reduction (MAR) algorithm based on a statistic iterative reconstruction (SIR) model for 2-D fan beam CT. Cubic spline interpolation is utilized to remove blocky black artifacts caused by incomplete projections, and it also makes the solution closer to the optimum. The dynamic wavelet thresholding method contains the benefits of both wavelet soft and hard thresholding methods and promotes the sparsity of the image, which is used to erase residual streaking artifacts and accelerate the convergence of the SIR. The algorithm is accelerated by graphics processing unit programming and costs 14.44s with 40 iterations, meeting the demand of clinical practice. The performance of the proposed algorithm is compared with two classical MAR algorithms: the total variation (TV) constraint SIR algorithm and the reweighted TV constraint SIR algorithm. The experimental datasets include nine artificial datasets. Both qualitative and quantitative evaluation results show the outstanding performance of the proposed algorithm in suppressing the metal artifacts and preserving the image details.

GPU based numerical simulation of core shooting process

Core shooting process is the most widely used technique to make sand cores and it plays an important role in the quality of sand cores. Although numerical simulation can hopefully optimize the core shooting process, research on numerical simulation of the core shooting process is very limited. Based on a two-fluid model (TFM) and a kinetic-friction constitutive correlation, a program for 3D numerical simulation of the core shooting process has been developed and achieved good agreements with in-situ experiments. To match the needs of engineering applications, a graphics processing unit (GPU) has also been used to improve the calculation efficiency. The parallel algorithm based on the Compute Unified Device Architecture (CUDA) platform can significantly decrease computing time by multi-threaded GPU. In this work, the program accelerated by CUDA parallelization method was developed and the accuracy of the calculations was ensured by comparing with in-situ experimental results photographed by a high-speed camera. The design and optimization of the parallel algorithm were discussed. The simulation result of a sand core test-piece indicated the improvement of the calculation efficiency by GPU. The developed program has also been validated by in-situ experiments with a transparent core-box, a high-speed camera, and a pressure measuring system. The computing time of the parallel program was reduced by nearly 95% while the simulation result was still quite consistent with experimental data. The GPU parallelization method can successfully solve the problem of low computational efficiency of the 3D sand shooting simulation program, and thus the developed GPU program is appropriate for engineering applications.

Parallel quantum-inspired evolutionary algorithms for community detection in social networks

The world around us may be viewed as a network of entities interconnected via their social, economic, and political interactions. These entities and their interactions form a social network. A social network is often modeled as a graph whose nodes represent entities, and edges represent interactions between these entities. These networks are characterized by the collective latent behavior that does not follow trivially from the behaviors of the individual entities in the network. One such behavior is the existence of hierarchy in the network structure, the sub-networks being popularly known as communities. Discovery of the community structure in a social network is a key problem in social network analysis as it refines our understanding of the social fabric. Not surprisingly, the problem of detecting communities in social networks has received substantial attention from the researchers.In this paper, we propose parallel implementations of recently proposed community detection algorithms that employ variants of the well-known quantum-inspired evolutionary algorithm (QIEA). Like any other evolutionary algorithm, a quantum-inspired evolutionary algorithm is also characterized by the representation of the individual, the evaluation function, and the population dynamics. However, individual bits called qubits, are in a superposition of states. As chromosomes evolve individually, the quantum-inspired evolutionary algorithms (QIEAs) are intrinsically suitable for parallelization.In recent years, programmable graphics processing units — GPUs, have evolved into massively parallel environments with tremendous computational power. NVIDIA® compute unified device architecture (CUDA®) technology, one of the leading general-purpose parallel computing architectures with hundreds of cores, can concurrently run thousands of computing threads. The paper proposes novel parallel implementations of quantum-inspired evolutionary algorithms in the field of community detection on CUDA-enabled GPUs.The proposed implementations employ a single-population fine-grained approach that is suited for massively parallel computations. In the proposed approach, each element of a chromosome is assigned to a separate thread. It is observed that the proposed algorithms perform significantly better than the benchmark algorithms. Further, the proposed parallel implementations achieve significant speedup over the serial versions. Due to the highly parallel nature of the proposed algorithms, an increase in the number of multiprocessors and GPU devices may lead to a further speedup.

LASSIE: simulating large-scale models of biochemical systems on GPUs

BackgroundMathematical modeling and in silico analysis are widely acknowledged as complementary tools to biological laboratory methods, to achieve a thorough understanding of emergent behaviors of cellular processes in both physiological and perturbed conditions. Though, the simulation of large-scale models—consisting in hundreds or thousands of reactions and molecular species—can rapidly overtake the capabilities of Central Processing Units (CPUs). The purpose of this work is to exploit alternative high-performance computing solutions, such as Graphics Processing Units (GPUs), to allow the investigation of these models at reduced computational costs.ResultsLASSIE is a “black-box” GPU-accelerated deterministic simulator, specifically designed for large-scale models and not requiring any expertise in mathematical modeling, simulation algorithms or GPU programming. Given a reaction-based model of a cellular process, LASSIE automatically generates the corresponding system of Ordinary Differential Equations (ODEs), assuming mass-action kinetics. The numerical solution of the ODEs is obtained by automatically switching between the Runge-Kutta-Fehlberg method in the absence of stiffness, and the Backward Differentiation Formulae of first order in presence of stiffness. The computational performance of LASSIE are assessed using a set of randomly generated synthetic reaction-based models of increasing size, ranging from 64 to 8192 reactions and species, and compared to a CPU-implementation of the LSODA numerical integration algorithm.ConclusionsLASSIE adopts a novel fine-grained parallelization strategy to distribute on the GPU cores all the calculations required to solve the system of ODEs. By virtue of this implementation, LASSIE achieves up to 92× speed-up with respect to LSODA, therefore reducing the running time from approximately 1 month down to 8 h to simulate models consisting in, for instance, four thousands of reactions and species. Notably, thanks to its smaller memory footprint, LASSIE is able to perform fast simulations of even larger models, whereby the tested CPU-implementation of LSODA failed to reach termination. LASSIE is therefore expected to make an important breakthrough in Systems Biology applications, for the execution of faster and in-depth computational analyses of large-scale models of complex biological systems.

Performance portability in reverse time migration and seismic modelling via OpenACC

Heterogeneity among the computational resources within a single machine has significantly increased in high performance computing to exploit the tremendous potential of graphics processing units (GPUs). Portability in terms of code development and performance has been a challenge due to major differences between GPU programming and memory models from one side and conventional central processing units (CPUs) from another side. Performance characteristics of compilers and processors also vary between machines. Emerging high-level directive-based programming models such as OpenACC has been proposed to target this challenge. In this work, we develop OpenACC implementations for both seismic modelling and reverse time migration algorithms that solve the isotropic, acoustic, and elastic wave equations. We employ OpenACC to take advantage of the computational power of two Nvidia GPU cards: (1) M2090 and (2) K40, residing in IBM and CRAY XC30 clusters respectively. We also explore the main aspects of hybridization seismic modelling and reverse time migration by implementing an Message Passing Interface (MPI)+OpenACC approach. We expose various mapping techniques to develop a portable code that maximizes performance regardless of compiler or platform. Depending on the intensity of the computations, different propagators exhibited different speedup behaviours against a full socket CPU MPI implementation. A performance enhancement of ~10× was obtained, when the acoustic model was ported to a single GPU, compared with a 1.7× speedup obtained using the isotropic model. Our MPI+OpenACC implementation of reverse time migration and seismic modelling shows promising scaling when multiple GPUs were used.

Detecting and counting people using real-time directional algorithms implemented by compute unified device architecture

This paper implements a real-time and directional counting algorithm using the Graphic Processing Unit (GPU) Programming for the purpose of detecting and counting people. We use the Compute Unified Device Architecture (CUDA) as the environment of the GPU programming. The proposed algorithm is implemented for detecting and counting people employing the single virtual line and two virtual lines, respectively, using three video streams and two GPU graphic cards GeForce GT 630 and GeForce GTX 550Ti. We first test the video streams on the algorithm by using GeForce GT 630 together with applying the single virtual line and two virtual lines, respectively. Then, we repeat the same procedures for the GPU graphic card GeForce GTX 550Ti. The obtained experimental results show that our proposed algorithm running on GPU can be successfully programmed and implemented for people detecting and counting problems.

GPU-accelerated Kernel Regression Reconstruction for Freehand 3D Ultrasound Imaging.

Volume reconstruction method plays an important role in improving reconstructed volumetric image quality for freehand three-dimensional (3D) ultrasound imaging. By utilizing the capability of programmable graphics processing unit (GPU), we can achieve a real-time incremental volume reconstruction at a speed of 25-50 frames per second (fps). After incremental reconstruction and visualization, hole-filling is performed on GPU to fill remaining empty voxels. However, traditional pixel nearest neighbor-based hole-filling fails to reconstruct volume with high image quality. On the contrary, the kernel regression provides an accurate volume reconstruction method for 3D ultrasound imaging but with the cost of heavy computational complexity. In this paper, a GPU-based fast kernel regression method is proposed for high-quality volume after the incremental reconstruction of freehand ultrasound. The experimental results show that improved image quality for speckle reduction and details preservation can be obtained with the parameter setting of kernel window size of [Formula: see text] and kernel bandwidth of 1.0. The computational performance of the proposed GPU-based method can be over 200 times faster than that on central processing unit (CPU), and the volume with size of 50 million voxels in our experiment can be reconstructed within 10 seconds.

Detecting Bank Conflict of GPU Programs Using Symbolic Execution—Case Study

GPU (Graphics Processing Unit) is used in various areas. Therefore, the demand for the verification of GPU programs is increasing. In this paper, we suggest the method to detect bank conflict by using symbolic execution. Bank conflict is one of the bugs happening in GPU and it leads the performance of programs lower. Bank conflict happens when some processing units in GPU access the same shared memory. Symbolic execution is the method to analysis programs with symbolic values. By using it, we can detect bank conflict on GPU programs which use many threads. We implement a prototype of the detector for bank conflict and evaluate it with some GPU programs. The result states that we can detect bank conflict on the programs with no loop regardless of the number of threads.

Survey of using GPU CUDA programming model in medical image analysis

Abstract With the technology development of medical industry, processing data is expanding rapidly and computation time also increases due to many factors like 3D, 4D treatment planning, the increasing sophistication of MRI pulse sequences and the growing complexity of algorithms. Graphics processing unit (GPU) addresses these problems and gives the solutions for using their features such as, high computation throughput, high memory bandwidth, support for floating-point arithmetic and low cost. Compute unified device architecture (CUDA) is a popular GPU programming model introduced by NVIDIA for parallel computing. This review paper briefly discusses the need of GPU CUDA computing in the medical image analysis. The GPU performances of existing algorithms are analyzed and the computational gain is discussed. A few open issues, hardware configurations and optimization principles of existing methods are discussed. This survey concludes the few optimization techniques with the medical imaging algorithms on GPU. Finally, limitation and future scope of GPU programming are discussed.

GPU acceleration of Monte Carlo simulations for polarized photon scattering in anisotropic turbid media.

In earlier studies, we developed scattering models and the corresponding CPU-based Monte Carlo simulation programs to study the behavior of polarized photons as they propagate through complex biological tissues. Studying the simulation results in high degrees of freedom that created a demand for massive simulation tasks. In this paper, we report a parallel implementation of the simulation program based on the compute unified device architecture running on a graphics processing unit (GPU). Different schemes for sphere-only simulations and sphere-cylinder mixture simulations were developed. Diverse optimizing methods were employed to achieve the best acceleration. The final-version GPU program is hundreds of times faster than the CPU version. Dependence of the performance on input parameters and precision were also studied. It is shown that using single precision in the GPU simulations results in very limited losses in accuracy. Consumer-level graphics cards, even those in laptop computers, are more cost-effective than scientific graphics cards for single-precision computation.

GSWO: A programming model for GPU-enabled parallelization of sliding window operations in image processing

Sliding Window Operations (SWOs) are widely used in image processing applications. They often have to be performed repeatedly across the target image, which can demand significant computing resources when processing large images with large windows. In applications in which real-time performance is essential, running these filters on a CPU often fails to deliver results within an acceptable timeframe. The emergence of sophisticated graphic processing units (GPUs) presents an opportunity to address this challenge. However, GPU programming requires a steep learning curve and is error-prone for novices, so the availability of a tool that can produce a GPU implementation automatically from the original CPU source code can provide an attractive means by which the GPU power can be harnessed effectively. This paper presents a GPU-enabled programming model, called GSWO, which can assist GPU novices by converting their SWO-based image processing applications from the original C/C++ source code to CUDA code in a highly automated manner. This model includes a new set of simple SWO pragmas to generate GPU kernels and to support effective GPU memory management. We have implemented this programming model based on a CPU-to-GPU translator (C2GPU). Evaluations have been performed on a number of typical SWO image filters and applications. The experimental results show that the GSWO model is capable of efficiently accelerating these applications, with improved applicability and a speed-up of performance compared to several leading CPU-to-GPU source-to-source translators.

Exploiting parallelism in the simulation of general purpose graphics processing unit program

The simulation is an important means of performance evaluation of the computer architecture. Nowadays, the serial simulation of general purpose graphics processing unit (GPGPU) architecture is the main bottleneck for the simulation speed. To address this issue, we propose the intra-kernel parallelization on a multicore processor and the inter-kernel parallelization on a multiple-machine platform. We apply these two methods to the GPGPU-sim simulator. The intra-kernel parallelization method firstly parallelizes the serial simulation of multiple compute units in one cycle. Then it parallelizes the timing and functional simulation to reduce the performance loss caused by the synchronization between different compute units. The inter-kernel parallelization method divides multiple kernels of a CUDA program into several groups and distributes these groups across multiple simulation hosts to perform the simulation. Experimental results show that the intra-kernel parallelization method achieves a speed-up of up to 12 with a maximum error rate of 0.009 4% on a 32-core machine, and the inter-kernel parallelization method can accelerate the simulation by a factor of up to 3.9 with a maximum error rate of 0.11% on four simulation hosts. The orthogonality between these two methods allows us to combine them together on multiple multi-core hosts to get further performance improvements.

RT-CUDA: A Software Tool for CUDA Code Restructuring

Recent development in graphic processing units (GPUs) has opened a new challenge in harnessing their computing power as a new general purpose computing paradigm. However, porting applications to CUDA remains a challenge to average programmers, which have to package code in separate functions, explicitly manage data transfers between the host and device memories, and manually optimize GPU memory utilization. In this paper, we propose a restructuring tool (RT-CUDA) that takes a C-like program and some user directives as compiler hints to produce an optimized CUDA code. The tool strategy is based on efficient management of the memory system to minimize data motion by managing the transfer between host and device, maximizing bandwidth for device memory accesses, and enhancing data locality and re-use of cached data using shared-memory and registers. Enhanced resource utilization is implemented by re-writing code as parametric kernels and use of efficient auto-tuning. The tool enables calling numerical libraries (CuBLAS, CuSPARSE, etc.) to help implement applications in science simulation like iterative linear algebra solvers. For the above applications, the tool implement an inter-block global synchronization which allow the execution overall among a few iterations which is helpful to balance load and to avoid polling. Evaluation of RT-CUDA has been performed using a variety of basic linear algebra operators (Madd, MM, MV, VV, etc.) as well as the programming of iterative solvers for systems of linear equations like Jacobi and Conjugate Gradient algorithms. Significant speedup has been achieved over other compilers like PGI OpenACC and GPGPU compilers for the above applications. Evaluation shows that generated kernels efficiently call math libraries and enable implementing complete iterative solvers. The tool help scientists developing parallel simulators like reservoir simulators, molecular dynamics, etc. without exposing to complexity of GPU and CUDA programming. We have partnership with a group of researchers at the Saudi Aramco, a national company in Saudi Arabia. RT-CUDA is currently explored as a potential development tool for applications involving linear algebra solvers by the above group. In addition, RT-CUDA is being used by Senior and Graduate students at King Fahd University of Petroleum and Minerals in their projects as part of RT-CUDA continuous enhancement.

FARGO3D: A NEW GPU-ORIENTED MHD CODE

ABSTRACT We present the FARGO3D code, recently publicly released. It is a magnetohydrodynamics code developed with special emphasis on the physics of protoplanetary disks and planet–disk interactions, and parallelized with MPI. The hydrodynamics algorithms are based on finite-difference upwind, dimensionally split methods. The magnetohydrodynamics algorithms consist of the constrained transport method to preserve the divergence-free property of the magnetic field to machine accuracy, coupled to a method of characteristics for the evaluation of electromotive forces and Lorentz forces. Orbital advection is implemented, and an N-body solver is included to simulate planets or stars interacting with the gas. We present our implementation in detail and present a number of widely known tests for comparison purposes. One strength of FARGO3D is that it can run on either graphical processing units (GPUs) or central processing units (CPUs), achieving large speed-up with respect to CPU cores. We describe our implementation choices, which allow a user with no prior knowledge of GPU programming to develop new routines for CPUs, and have them translated automatically for GPUs.

Real-time Machine Vision with GPU-acceleration using Quasar

The computational performance of graphical processing units (GPUs) has improved significantly, achieving even speed-up factors of 10x-50x compared to single-threaded CPU execution are not uncommon. This makes their use for high throughput machine vision very appealing. However, GPU programming is challenging, requiring a significant programming expertise. We present a new programming framework that mitigates the challenges common for GPU programming while maintaining the significant acceleration.

A GPU-based parallel Object kinetic Monte Carlo algorithm for the evolution of defects in irradiated materials

In this work we present a parallel Object kinetic Monte Carlo (OkMC) computational model implemented using GPUs (graphics processing units) computing to simulate the evolution of interacting random walkers in materials. Different test simulations were performed under different conditions and compared to analytical solutions and sequential OkMC codes. The simulation results obtained with our algorithm show to be in excellent agreement with sequential OkMC codes and analytical solutions. The speedups reached with the GPU programming with respect to sequential OkMC codes in the test cases used here were in the range [30–100]. This increased efficiency allows to follow the evolution of millions of interacting particles in a relatively short computational time in contrast to classical OkMC codes. We expect this parallel kMC algorithm based on GPU programming to allow for investigations on defect evolution in materials using simulation boxes of realistic dimensions and physical times close to those achieved experimentally.

A collaborative client participant fusion system for realistic remote conferences

Remote conferencing systems provide a shared environment where people in different locations can communicate and collaborate in real time. Currently, remote video conferencing systems present separate video images of the individual participants. To achieve a more realistic conference experience, we enhance video conferencing by integrating the remote images into a shared virtual environment. This paper proposes a collaborative client participant fusion system using a real-time foreground segmentation method. In each client system, the foreground pixels are extracted from the participant images using a feedback background modeling method. Because the segmentation results often contain noise and holes caused by adverse environmental lighting conditions and substandard camera resolution, a Markov Random Field model is applied in the morphological operations of dilation and erosion. This foreground segmentation refining process is implemented using graphics processing unit programming, to facilitate real-time image processing. Subsequently, segmented foreground pixels are transmitted to a server, which fuses the remote images of the participants into a shared virtual environment. The fused conference scene is represented by a realistic holographic projection.

GPGPU-MiniBench: Accelerating GPGPU Micro-Architecture Simulation

Graphics processing units (GPU), due to their massive computational power with up to thousands of concurrent threads and general-purpose GPU (GPGPU) programming models such as CUDA and OpenCL, have opened up new opportunities for speeding up general-purpose parallel applications. Unfortunately, pre-silicon architectural simulation of modern-day GPGPU architectures and workloads is extremely time-consuming. This paper addresses the GPGPU simulation challenge by proposing a framework, called GPGPU-MiniBench, for generating miniature, yet representative GPGPU workloads. GPGPU-MiniBench first summarizes the inherent execution behavior of existing GPGPU workloads in a profile. The central component in the profile is the Divergence Flow Statistics Graph (DFSG), which characterizes the dynamic control flow behavior including loops and branches of a GPGPU kernel. GPGPU-MiniBench generates a synthetic miniature GPGPU kernel that exhibits similar execution characteristics as the original workload, yet its execution time is much shorter thereby dramatically speeding up architectural simulation. Our experimental results show that GPGPU-MiniBench can speed up GPGPU architectural simulation by a factor of 49 $\times$ on average and up to 589 $\times$ , with an average IPC error of 4.7 percent across a broad set of GPGPU benchmarks from the CUDA SDK, Rodinia and Parboil benchmark suites. We also demonstrate the usefulness of GPGPU-MiniBench for driving GPU architecture exploration.

Real-time single camera natural user interface engine development

Natural user interfaces (NUIs) provide human computer interaction (HCI) with natural and intuitive operation interfaces, such as using human gestures and voice. We have developed a real-time NUI engine architecture using a web camera as a means of implementing NUI applications. The system captures video via the web camera, implements real-time image processing using graphic processing unit (GPU) programming. This paper describes the architecture of the engine and the real-virtual environment interaction methods, such as foreground segmentation and hand gesture recognition. These methods are implemented using GPU programming in order to realize real-time image processing for HCI. To verify the efficacy of our proposed NUI engine, we utilized it in the development and implementation of several mixed reality games and touch-less operation applications, using the developed NUI engine and the DirectX SDK. Our results confirm that the methods implemented by the engine operate in real time and the interactive operations are intuitive.

Accelerated Molecular Mechanical and Solvation Energetics on Multicore CPUs and Manycore GPUs.

Despite several reported acceleration successes of programmable GPUs (Graphics Processing Units) for molecular modeling and simulation tools, the general focus has been on fast computation with small molecules. This was primarily due to the limited memory size on the GPU. Moreover simultaneous use of CPU and GPU cores for a single kernel execution - a necessity for achieving high parallelism - has also not been fully considered. We present fast computation methods for molecular mechanical (Lennard-Jones and Coulombic) and generalized Born solvation energetics which run on commodity multicore CPUs and manycore GPUs. The key idea is to trade off accuracy of pairwise, long-range atomistic energetics for higher speed of execution. A simple yet efficient CUDA kernel for GPU acceleration is presented which ensures high arithmetic intensity and memory efficiency. Our CUDA kernel uses a cache-friendly, recursive and linear-space octree data structure to handle very large molecular structures with up to several million atoms. Based on this CUDA kernel, we present a hybrid method which simultaneously exploits both CPU and GPU cores to provide the best performance based on selected parameters of the approximation scheme. Our CUDA kernels achieve more than two orders of magnitude speedup over serial computation for many of the molecular energetics terms. The hybrid method is shown to be able to achieve the best performance for all values of the approximation parameter. The source code and binaries are freely available as PMEOPA (Parallel Molecular Energetic using Octree Pairwise Approximation) and downloadable from http://cvcweb.ices.utexas.edu/software.