AMD DirectGMA

Introduction

DirectGMA is AMD’s proprietary API for direct memory transfers between PCI devices. This technology enables very high performance, very low latency real time processing. Using DirectGMA, BitFlow frame grabbers are able to acquire images directly to the memory of AMD boards, allowing immediate GPU processing using OpenCL kernels.

Dependencies

  • The AMD APP SDK
  • AMD DirectGMA SDK
  • An AMD FirePro (or other DirectGMA compatible) GPU with DirectGMA enabled
  • The BitFlow SDK versions 6.10 or greater
  • A modern BitFlow frame grabber compatible with the BiAPI (eg., Alta, Neon, Karbon, Cyton, Axion)
  • Microsoft Visual Studio 2012 or greater
  • Examples

SimpleDGMA

Performs an BufIn acquisition with DirectGMA, processing each frame with a simple OpenCL image gain kernel, and displaying the resultant frames to a GLUT preview window.