Authors: Philipp Lucas
Addresses: Compiler Design Lab, Saarland University, Saarbrucken, Germany
Abstract: Modern Graphics Processing Units (GPUs) commonly found in today|s PCs feature multiple processing units and can be used for general purpose computations and in particular, parallel numerical algorithms. But the available level of abstraction is still very low. Typically, GPU programs are written in assembly language. In this paper, the architecture which is still tightly coupled to the rasterisation algorithm that GPUs are originally meant to implement, and some of the algorithms efficiently implemented on GPUs so far, will be presented. The strengths and weaknesses of GPUs and approaches towards the goal of an easily programmable GPU are presented.
Keywords: SIMD; GPU; vector processor; graphics hardware; ubiquitous parallelism; parallel computing; scientific computation; graphics processing units.
International Journal of Computational Science and Engineering, 2005 Vol.1 No.2/3/4, pp.142 - 156
Published online: 05 May 2006 *Full-text access for editors Access for subscribers Purchase this article Comment on this article