Hands-On High Performance with Go

Docker for GPU-enabled programming

If you'd like to use Docker for your GPU-enabled programming in this chapter, you can perform the following steps, but to use this, you must have a compatible NVIDIA CUDA GPU in your computer. You can find a full list of enabled cards at https://developer.nvidia.com/cuda-gpus.

We might not use Docker in this way in a production environment for GPU-accelerated computing, because you'd most likely want to be as close to the hardware as possible for GPU-accelerated programming, but I've chosen to use this methodology in this chapter in order to have a reproducible build for the consumer of this book to use. Most of the time a reproducible build is an acceptable trade off for the slight performance penalty of ...