high performance computing on graphics processing units: hgpu.org

This paper introduces Sample Distribution Shadow Maps (SDSMs), a new algorithm for hard and soft-edged shadows that greatly reduces undersampling, oversampling, and geometric aliasing errors compared to other shadow map techniques. SDSMs fall into the space between scene-dependent, variable-performance shadow algorithms and scene-independent, fixed-performance shadow algorithms. They provide a fully automated solution to shadow map […]

Thanks to the advances in parallel processing hardware, iterative algorithms for cone beam reconstruction are now available with computation times acceptable for clinical use. At the same time they are able to accomodate more accurately the physical effects underlying the X-Ray imaging process. Many parameters are involved, which need to be precisely calibrated in order […]

In this paper, we present the development of a highspeed volume rendering system that combines 3D texture compression and parallel programming techniques for rendering multiple high-resolution 3D images obtained with medical or industrial CT. The 3D texture compression algorithm (DXT5) provides extremely high efficiency since it reduces the memory consumption to 1/4 of the original […]

We study the use of a GPU for the numerical approximation of the curvature dependent flows of graphs – the mean-curvature flow and the Willmore flow. Both problems are often applied in image processing where fast solvers are required. We approximate these problems using the complementary finite volume method combined with the method of lines. […]

We demonstrate real time off-axis Quantitative Phase Imaging (QPI) using a phase reconstruction algorithm based on NVIDIA’s CUDA programming model. The phase unwrapping component is based on Goldstein’s algorithm. By mapping the process of extracting phase information and unwrapping to GPU, we are able to speed up the whole procedure by more than 18.8x with […]

Extracting performance improvements from modest and cost-effective computing resources is one of the key challenges in the IT sector. CPU clock speeds have reached a plateau in recent years, with no significant clock speed improvements forthcoming. However, we see an increasing number of computational cores available on the desktop, via the CPU and, more recently, […]

Optical flow estimation (the estimation of the apparent motion of objects in an image sequence) is used in many applications like video compression, object detection and tracking, robot navigation, and so on. This project was focussed on one specific optical flow estimation algorithm, which uses directional filters and an AM-FM demodulation algorithm for the estimation […]

This paper describes an optical flow estimation algorithm using directional filters and an AM-FM demodulation algorithm, and its efficient implementation on a NVIDIA GPU using CUDA. The resulting implementation is several thousand times faster than the corresponding MATLAB code, which makes the described scheme suitable for real-time applications. This paper also describes a new multiresolution […]

The implementation via CUDA of a hybrid dense dynamic programming method for knapsack problems on amulti-GPU architecture is considered. Tests are carried out on a Bull cluster with Tesla S1070 computing systems. A first series of computational results shows substantial speedup. The speedup factor is close to 28 with two GPUs.

A number of issues related to floating point accuracy and compliance are a frequent source of confusion on both CPUs and GPUs. The purpose of this white paper is to discuss the most common issues related to NVIDIA GPUs and to supplement the documentation in the CUDA C Programming Guide.

This paper proposes and describes a methodology developed to port complex scientific applications originally written in FORTRAN to nVidia CUDA. The significance of this lies in the fact that, despite the performance improvement and programmer-friendliness provided by CUDA, it presently lacks support for FORTRAN. The methodology described in this paper addresses this problem using a […]

Finding interesting celestial objects among tens of thousands or even millions of recorded raw data is not an easy task to implement. In this paper we speed up this process with high level nvidia cuda C++ template library called Thrust, which makes our database with R interface much more evaluatedcient.