Computation on GPUs: From a Programmable Pipeline to an Efficient Stream Processor

@article{Comba2003ComputationOG,
  title={Computation on GPUs: From a Programmable Pipeline to an Efficient Stream Processor},
  author={Jo{\~a}o Luiz Dihl Comba and Carlos A. Dietrich and Christian Azambuja Pagot and Carlos Eduardo Scheidegger},
  journal={RITA},
  year={2003},
  volume={10},
  pages={41-70}
}
The recent development of graphics hardware is presenting a change in the implementation of the graphics pipeline, from a fixed set of functions, to userdeveloped special programs to be executed on a per-vertex or per-fragment basis. This programmability allows the efficient implementation of different algorithms directly on the graphics hardware. In this tutorial we will present the main techniques that are involved in implementing algorithms in this fashion. We use several test cases based on… CONTINUE READING