Machine learning apps in containers can’t run GPU-accelerated code, but a new Docker plugin by Nvidia is set to remedy all that.
Nvidia, developer of the CUDA standard for GPU-accelerated programming, is releasing a plugin for the Docker ecosystem that makes GPU-accelerated computing possible in containers. With the plugin, applications running in a Docker container get controlled access to the GPU on the underlying hardware via Docker’s own plug-in system.
Read more at InfoWorld
