skills.rest
coreweave
coreweaveOFFICIAL
Explore all skills by this author →
WebsiteGitHub

add-gpu-nodes-to-gke

Automate GPU node pool creation, driver setup, and Slurm configuration on GKE.

7|Updated Jul 9, 2026
Search Tags:#GKE#Slurm#SUNK#GPU node pool#GKE SUNK deployment
One-click install
npx skills add https://github.com/coreweave/sunk-anywhere --skill add-gpu-nodes-to-gke
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: add-gpu-nodes-to-gke
Source: https://github.com/coreweave/sunk-anywhere/tree/main/skills/gke/add-gpu-nodes-to-gke
Command: npx skills add https://github.com/coreweave/sunk-anywhere --skill add-gpu-nodes-to-gke

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires gcloud, helm, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill automates the process of adding GPU compute nodes to an existing SUNK (Slurm on Kubernetes) deployment running on GKE. It handles node pool creation, driver setup, Slurm configuration, and GPU job verification.

Core Features & Use Cases

  • Node Pool Creation: Automates the creation of GKE GPU node pools with driver installation.
  • NVIDIA Driver Path Setup: Configures NVIDIA driver paths for slurmd processes.
  • Slurm GPU Workers: Defines Slurm GPU workers and patches GRES configuration.
  • GPU Job Verification: Validates GPU job execution with smoke tests.
  • Use Case: Ideal for enhancing a SUNK on GKE deployment with GPU capabilities for running GPU jobs and executing CUDA-based applications.

Quick Start

Use the 'add-gpu-nodes-to-gke' skill to add GPU nodes to your GKE SUNK deployment.

Frequently Asked Questions about add-gpu-nodes-to-gke

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I add GPU nodes to a Slurm on Kubernetes deployment running on GKE?▼

To add GPU compute nodes to a SUNK deployment on GKE, you must create a GPU node pool, configure NVIDIA driver paths, patch Slurm GRES configuration, and verify CUDA execution. This process is automated using gcloud and Helm.

What is needed to automate GKE GPU node pool creation for Slurm workers?▼

You need an existing GKE cluster and a pre-installed SUNK deployment. Additionally, the Google Cloud SDK (gcloud) and Helm must be installed to orchestrate the GPU node pool creation and driver setup.

How does Slurm GPU configuration handle NVIDIA driver paths for slurmd processes?▼

The configuration defines Slurm GPU workers and patches GRES settings to ensure slurmd processes recognize the hardware. It explicitly configures NVIDIA driver paths so Slurm can correctly allocate GPU resources to CUDA jobs.

Can I run CUDA-based applications on SUNK after adding a GKE GPU node pool?▼

Yes, adding a GKE GPU node pool to SUNK enables CUDA-based application execution. The automation includes GPU job verification using smoke tests to validate that your Slurm workers can successfully execute CUDA workloads.

What are the limitations of automating GKE GPU node addition for SUNK?▼

The main limitation is that it requires an existing GKE cluster and a pre-existing SUNK deployment. It cannot initialize a new Slurm on Kubernetes environment, but strictly manages adding GPU compute nodes, driver setup, and Slurm configuration.

Related Skills

Explore More Software Engineering Skills→
coreweavecoreweave
7

deploy-sunk-on-gke

Automate SUNK deployment on Google Kubernetes Engine with cluster and Slurm setup.

Official
Advanced
niznik-devniznik-dev
16

Run Experiment

Orchestrate SLURM fine-tuning jobs with torchtune and evaluate with inspect-ai.

Community
Advanced
ssmleossmleo

cloud-run-basics

Automate deployment and management of Cloud Run services, jobs, and worker pools.

Community
Advanced
devendrapratapsinghdevendrapratapsingh
1

orchestration

Design, deploy, and manage Kubernetes clusters and container orchestration platforms.

Community
Advanced