asdf-hashicorp VS aws-virtual-gpu-device-plugin

brew install asdf asdf plugin-add terraform https://github.com/asdf-community/asdf-hashicorp.git asdf install terraform 0.12 asdf exec terraform destroy

I'm a huge fan of asdf and have been using for years together with direnv! It's great to see how much effort is put into it! I hope more people adopt it so that we don't have to `curl | sh`! One thing I have issues with asdf is security as are no checksums, so, you if I project get compromised you'll get compromised, too. This, of course, is in addition to the third-party asdf plugin getting itself compromised (which is the greater risk). Last, but not least - I wish asdf came with something like eget [0] incorporated so that it can install 99% of the plugins directly and safely! Last, but not least - 99% of the plugins have almost identical code and all that changes is the repo, so, this should be generalized. For example, many years ago I made just one codebase of all HashiCorp plugins [1] and it's been working great!

[0]: https://github.com/zyedidia/eget

[1]: https://github.com/asdf-community/asdf-hashicorp

CMD> asdf plugin list all Output sample: [...] kubebuilder https://github.com/virtualstaticvoid/asdf-kubebuilder.git kube-capacity https://github.com/looztra/asdf-kube-capacity.git kubectl *https://github.com/Banno/asdf-kubectl.git kubectl-bindrole https://github.com/looztra/asdf-kubectl-bindrole.git kubectx *https://gitlab.com/wt0f/asdf-kubectx.git kubefedctl https://github.com/kvokka/asdf-kubefedctl.git kubeseal https://github.com/stefansedich/asdf-kubeseal.git kubesec https://github.com/vitalis/asdf-kubesec.git kubespy https://github.com/jfreeland/asdf-kubespy.git kubeval https://github.com/stefansedich/asdf-kubeval.git terraform *https://github.com/Banno/asdf-hashicorp.git terraform-docs *https://github.com/looztra/asdf-terraform-docs.git terraform-lsp https://github.com/bartlomiejdanek/terraform-lsp.git terraform-validator https://github.com/looztra/asdf-terraform-validator.git terragrunt https://github.com/lotia/asdf-terragrunt.git [...]

I've been using it and has contributed to it, but I wish it supported out of the box GitHub-based release binaries as 90% of my code for different vendors was relatively the same, so, instead of having multiple identical repositories, I created one, which uses introspection [0].

I wish this was available out of the box to handle literary 90% of the tools.

Also, I typically pair it with direnv [1] for even more magic.

[0]: https://github.com/Banno/asdf-hashicorp

[1]: https://direnv.net/

This project (available here) uses the k8s device plugin described by this AWS blog post to make GPU-based nodes publish the amount of GPU resource they have available. Instead of the amount of VRAM available or some abstract metric, this plugin advertises the amount of pods/processes that can be connected to the GPU. This is controlled by what is called by NVIDIA as Multi-Process Service (MPS).

AWS has apparently already started using this type of tech as of this year (see lost below). They mention virtual gpus but this particular solution probably won't help OP unfortunately. https://aws.amazon.com/blogs/opensource/virtual-gpu-device-plugin-for-inference-workload-in-kubernetes/

The post explores GPU device plugin to address how to set fractional number of GPU resource for each pod by implementing the Kubernetes device plugin and Nvidia MPS. This project has been open sourced on GitHub.