Wednesday, May 25, 2016

Episode 54: Programming for the GPU

Download

We are sponsored by audible! http://www.audibletrial.com/programmingthrowdown

We are on Patreon! https://www.patreon.com/programmingthrowdown


Programming for the GPU with Mark Harris


  • Introductions: Who is Mark Harris? What is NVIDIA?
  • GPUs & Heterogeneous Computing
    • What's the performance difference between GPU and CPU?
    • What’s better on a GPU? What’s better on a CPU?
    • If a program needs to use the CPU (e.g. reading packets from the network) and the GPU (e.g. fluid simulation) at the same time, how do the CPU and GPU programs communicate?
  • CUDA
    • What is CUDA?
    • What’s the difference between SSE / CUDA / OpenCL?
    • How does someone debug CUDA code?
    • How does someone profile their code to find bottlenecks?
    • Cuda Plattorm
      • Directives, openmp
      • Cuda c++
  • CUDA Libraries
    • Deep Learning
    • CuBLAS
    • Drive
  • Deep Dream:
    • https://github.com/google/deepdream
  • Udacity course on CUDA
    • Intro to parallel programming
  • Life @ NVIDIA
    • What is a day working at NVIDIA like?