Skip to content
Hardware for Neural Networks (2017 Edition) - beamer slides
TeX
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
Double_Floating_Point_Format.png
Fermi_memory_hierarchy.png
Fermi_memory_hierarchy_speed_size.png
GeForce_GTX_1080_Block_Diagram_FINAL.png
GeForce_GTX_1080_SM_Diagram_FINAL.png
Half_Floating_Point_Format.png
IMGPipeline.png
README.md
slides.tex

README.md

Hardware for Neural Networks (2017 Edition)

I was just curious about current hardware options for neural networks and ended up giving a talk.

Brief affair with LaTex and Beamer - passion still strong after all those years :-)

  • Is GPU really faster? (How much?)
  • Which parameter is important for performance?
  • Why is only NVidia ever mentioned as an option?
  • Is there anything interesting expected for 2017?

Hopefully no one mind me using all those images from web. If so please contact me.

You can’t perform that action at this time.