Most data-intensive operations for multimedia applications such as image processing, vision, and 3D graphics require high external memory bandwidth. In augmented-reality (AR) processors [1], both 3D graphics and vision operations are required, so memory bandwidth becomes even more critical. In [1], however, memory bandwidth is not considered, floating-point processing is not supported, and ther...

The AMD Fusion Family of Accelerated Processing Units (APUs), introduced to market in January 2011, is a new generation of processors that combines the computing processing unit (CPU) and graphics processing unit (GPU) capabilities in a single chip (die). APU-based platforms can deliver a prodigious amount of computational horsepower, and can present enormous opportunities in developing an appl...

Thanks to their high performance and programmability, the latest graphics cards can now be used for scientific purpose. They are indeed very efficient parallel Single Instruction Multiple Data (SIMD) machines. This new trend is called General Purpose computation on Graphics Processing Unit (GPGPU [13]). Regarding the stereo problem, variational methods based on deformable models provide dense, ...

The implementation of a 2D Euler solver on graphics hardware is described. The graphics processing unit is highly parallelised and uses a programming model that is well suited to flow computation. Results for a transonic turbine cascade test-case are presented. For large grids (10 nodes) a 40 times speed-up compared to a Fortran implementation on a contemporary CPU is observed.

The increasing array size of radio astronomy interferometers is causing the associated computation to scale quadratically with the number of array signals. Consequently, efficient usage of alternate processing architectures should be explored in order to meet this computational challenge. Affordable parallel processors have been made available to the general scientific community in the form of ...

Recently, the use of graphics processing units as a means of achieving the hardware acceleration of the finite-difference time-domain (FDTD) technique has attracted significant interest in the computational electromagnetics community. However, the large memory requirements of the FDTD, compounded by the limited memory resources available in graphics processing units, compromise the efficiency o...