The performance portability of OpenCL kernel implementations for common memory bandwidth limited linear algebra operations across different hardware generations of the same vendor as well as across vendors is studied. Certain combinations of kernel implementations and work sizes are found to exhibit good performance across compute kernels, hardware generations, and, to a lesser degree, vendors. […]

September 3, 2014 by hgpu

Frequent pattern mining is one of the widely used data mining techniques for discovering trends or patterns from databases. As data is growing in exponential pace, data mining activities need more powerful computing. Fortunately modern GPUs (Graphics Processing Units) have specialized electronic circuits and support parallel processing. GPUs are capable of processing huge amount of […]

August 23, 2014 by hgpu

In this work we propose an efficient finite volume approximation of two-fluid flows. Our scheme is based on three ingredients. We first construct a conservative scheme that removes the pressure oscillations phenomenon at the interface. The construction relies on a random sampling at the interface [6, 5]. Secondly, we replace the exact Riemann solver by […]

March 15, 2014 by hgpu

Graphics processing units (GPUs) in a computer system are increasingly adopted to create a realistic environment in 3D applications. Despite the extremely high parallelism of these devices with a tremendous amount of processing elements, GPUs are seldom used in scientific applications owing to their difficulty in programming. Additionally, vendor-provided APIs are too specific to cross […]

December 9, 2013 by hgpu

In this paper, we propose a new very simple numerical method for solving liquid-gas compressible flows. Such flows are difficult to simulate because classic conservative finite volume schemes generate pressure oscillations at the liquid-gas interface. We extend to several dimensions the random choice scheme that we have constructed in [2]. The extension is performed through […]

July 22, 2013 by hgpu

Heterogeneous multicore architectures with CPU and add-on GPUs or streaming processors are now widely used in computer systems. These GPUs provide substantially more computation capability and memory bandwidth compared to traditional multi-cores. Also, because they are highly programmable, they provide the computational performance needed for realistic graphics rendering. Applications with general computations can also be […]

July 19, 2013 by hgpu

High-performance computing systems today include a variety of compute devices such as multi-core CPUs, GPUs and many-core accelerators. OpenCL allows programming different types of compute devices using a single API and kernel language. However, there is no standard matrix operations library in OpenCL for operations such as matrix multiplication that works well on a variety […]

May 11, 2013 by hgpu

Reflector antennas play a key role in the communication industry, and enhancing the speed of the analysis of reflector antenna systems can provide better responsiveness to the needs of industry as well as promote better understanding of software modeling through faster visualization. A reflector antenna system typically consists of a feed assembly, with a feedhorn […]

May 6, 2013 by hgpu

The main goal of this project is to develop a computer program to model the creation of structure in Saturn’s ring system. The computer program will be used to answer these questions: (1) How are gaps in Saturn’s Rings formed; (2) how accurately can I model gap formation with a 3D N-Body simulation; and (3) […]

April 21, 2013 by hgpu

Volume rendering is quite an old concept of representing images, dating back to the 1980’s. It is very useful in the medical field for visualizing the results of a computer tomography (CT) and magnet resonance tomography (MRT) in 3D. Apart from these two major applications for volume rendering, there aren’t many other fields of usage […]

March 18, 2013 by hgpu

We present several numerical simulations of conservation laws on recent multicore processors, such as GPU’s, using the OpenCL programming framework. Depending on the chosen numerical method, different implementation strategies have to be considered, for achieving the best performance. We explain how to program efficiently three methods: a finite volume approach on a structured grid, a […]

December 2, 2012 by hgpu

In this paper, we propose a new very simple numerical method for solving liquid-gas compressible flows. Such flows are difficult to simulate because classical conservative finite volume schemes generate pressure oscillations at the liquid-gas interface. We extend to several dimensions the random choice scheme that we have previously constructed. The extension is performed through Strang […]

December 1, 2012 by hgpu