Skip to content
Collection of tools and examples for managing Accelerated workloads in Kubernetes Engine
Go Shell Python Makefile Dockerfile
Branch: overcommit
Clone or download
Pull request Compare This branch is 6 commits ahead, 6 commits behind GoogleCloudPlatform:master.
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.
build
cmd/nvidia_gpu
demo
example
nvidia-driver-installer
pkg/gpu/nvidia
vendor
.gitignore
.travis.yml
CONTRIBUTING.md
Dockerfile
Gopkg.lock
Gopkg.toml
LICENSE
Makefile
README.md
VERSION
daemonset.yaml

README.md

nvidia-gpu-device-plugin with GPU overcommitmemt

This repository is a fork of the nvidia-gpu-device-plugin in Google Kubernetes Engine. This fork allows overcommitment of GPU resources, meant so multiple pods can share 1 GPU by reporting more actual available GPUs to Kubernetes. This can be set by setting -overcommit-multiply-by=2 or a higher number, this specifies how many times 1 GPU will be reported.

More details on the nvidia-gpu-device-plugin are here.

You can’t perform that action at this time.