GPUs are competing FPGA-based image enhancement for specialized cameras

• Tweet

• Tweet

Above is a short lecture by Dr. Reinhard Borst,  ELTEC Elektronik AG. Lecture has been delivered at SpectroNet Collaboration Forum Konstanz, 2012.

Slides are available as a PDF. Lecture key points:

• FPGAs for longevity
  • FPGA are long-living
  • Programming in VHDL
• GPUs for best performance
  • 3 competing approaches (AMD, Intel, Nvidia)
  • Consumer-driven, fast-changing
  • Under OpenCL or CUDA (proprietary)
  • GPUs come for „free“ – but who can program them?
  • GPUs become HOT
• FPGA for industrial – GPU for commercial?

Tags: FPGA, GPU, image processing

Category: Computer Science, Video