Implementing common array algorithms on modern GPUs

  • Authors:
  • Tzvetomir Vassilev

  • Affiliations:
  • -

  • Venue:
  • CompSysTech '09 Proceedings of the International Conference on Computer Systems and Technologies and Workshop for PhD Students in Computing
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper proposes GPU implementations of common array algorithms: minimum/maximum, sum and count of elements that meet a certain condition. The first approach utilises the reduction-type operations on the GPU, but the second technique is based on GPGPU programming plus hardware occlusion queries. The testing programs were implemented using OpenGL and GL shading language (GLSL) for programming the GPU, which makes the approach rather general. The results section shows a huge performance advantage of this approach compared to the traditional one based on the CPU implementation.