General purpose GPU programming

Short Description
General purpose GPU programming. Dimitar Atanasov. Abstract: The paper describes programming features of the modern graphics hardware. It accents on …

Website: ecet.ecs.ru.acad.bg | Filesize: 103kb

Content
International Conference on Computer Systems and Technologies - CompSysTech’ 2005
- 1 -
General purpose GPU programming
Dimitar Atanasov
Abstract: The paper describes programming features of the modern graphics hardware. It accents on
the features introduced in Microsoft DirectX 9.0 and using them in general purpose programming of the
graphics hardware for effective offload of the CPU. As reference is used NVIDIA’s GeForce FX GPUs.
Key words: GPU, DirectX, Pixel and Vertex Shaders.
INTRODUCTION
Modern computer graphics hardware contain extremely powerful graphics
processing units (GPU). These GPUs are designed to perform a limited number of
operations on very large amounts of data. They typically have more than one processing
pipeline working in parallel with each other. They can in fact be thought of as highly
parallel Single Instruction Multiple Data (SIMD) type processors.
The performance of these GPUs is also growing at an extraordinary rate. In fact
over the last decade or so the processing power of GPUs has been growing at a rate
faster than Moore’s Law, which governs the performance growth rate of CPUs. In a
presentation at Graphics Hardware 2003 titled “Data Parallel Computing on Graphics
Hardware”, Ian Buck estimated that the current Nvidia…