high performance computing on graphics processing units: hgpu.org

We present a study of three important kernels that occur frequently in iterative statistical applications: K-Means, Multi-Dimensional Scaling (MDS), and PageRank. We implemented each kernel using OpenCL and evaluated their performance on an NVIDIA Tesla GPGPU card. By examining the underlying algorithms and empirically measuring the performance of various components of the kernel we explored […]

Large heterogeneous data centers of today lack methods to appraise the best fitting solutions regarding, among others, hardware acquisition cost, development time, and performance. Especially resource intensive applications benefit from increased data center utilization to leverage heterogeneous resources and accelerators. In this paper, we implement various methods to accelerate a seismic modeling application, which is […]

Over recent years, using Graphics Processing Units (GPUs) has become as an effective method for increasing the performance of many applications. However, these performance benefits from GPUs come at a price. Firstly extensive programming expertise and intimate knowledge of the underlying hardware are essential for gaining good speedups. Secondly, the expressibility of GPU-based programs are […]

The future multi-modal user interfaces of battery-powered mobile devices are expected to require computationally costly image analysis techniques. The use of Graphic Processing Units for computing is very well suited for parallel processing and the addition of programmable stages and high precision arithmetic provide for opportunities to implement energy-efficient complete algorithms. At the moment the […]

This project was done with the Mobile Microprocessor Group at Intel Corporation as a part of a six month internship. The primay objective of this project was to study the performance of GPGPUs (General purpose computation on Graphics Processing Units) for various benchmark applications. GPGPUs have gained wide spread importance in recent years because of […]

This thesis presents an analysis of numerical accuracy issues that are found in many scientific GPU applications due to floating-point computation. Two widely held myths about floating-point on GPUs are that the CPU’s answer is more precise than the GPU version and that computations on the GPU are unavoidably different from the same computations on […]

The test and evaluation community faces conflicting pressures: Provide more computing power and reduce electrical power requirements, both on the range and in the laboratory. The authors present some quantifiable benefits from the implementation of General Purpose Graphics Processing Units (GPGPUs) as heterogeneous processors. This produces power, space, cooling, and maintenance benefits that they have […]

Parallel projection based Single Photon Emission Computed Tomography (SPECT) is one of the most widely used nuclear imaging technique even nowadays. Serious artefacts are produced in the reconstructed images due to the non-homogeneous attenuation medium and the distance dependent spatial resolution (DDSR) of the parallel imaging. Effective non-uniform attenuation correction and DDSR reduction procedures should […]

Feasibility studies into the use of GPUs have been performed on two key algorithms in the ATLAS High Level Trigger. A GPU-based version of the Z-finder routine was found to give up to 35 times speedup in the best case scenario, while a speed-up of over 5 times was observed in a GPU-based Kalman Filter […]

High performance and relatively low cost of GPU-based platforms provide an attractive alternative for general purpose high performance computing (HPC). However, the emerging HPC applications have usually stricter output correctness requirements than typical GPU applications (i.e., 3D graphics). This paper first analyzes the error resiliency of GPGPU platforms using a fault injection tool we have […]

Modern graphics accelerators do not serve for classic computer games graphics computation accelerations only any more. Their highly parallel architectures enable their use in a broad spectrum of calculations. Because of the release of the OpenCL library and our interest in ray-tracing, we decided to show that ray-tracing is feasible not only on a multi-core […]

We present a portable OpenCL implementation of the radix sort algorithm. We test it on several GPUs or CPUs in order to assess its good performances on different hardware. We also apply our implementation to the Particle-In-Cell (PIC) sorting, which is useful in plasma physics simulations.