Skip to main content

Deploy GPU VMs using Terraform

Manage GPU VMs using nodeshift_bucket resource:

resource "nodeshift_gpu" "example" {
  gpu_name = "RTX_3090"
  image = "ubuntu:latest"
  region = "Europe"
  ssh_key = "ssh-rsa ..."
  gpu_count = 1
}

Parameters​

Required​

  • gpu_name (String) Name of the GPU to be used in deployment
  • image (String) Image used to run your Docker container name:version
  • ssh_key (String) SSH key to add to the target GPU to make it possible to connect to your deployment

Optional​

  • gpu_count (Number) Number of GPU instances for your deployment
  • region (String) Region where you want to deploy your GPU.

Read-Only​

  • uuid (String) String UUID of the GPU, computed

Available options for region parameter:

  • Northern America
  • Central America
  • South America
  • Europe
  • Asia
  • Africa
  • Oceania
  • Caribbean

Available list of gpus to use in gpu_name:

GPU Model Name
A100 PCIE
A100 SXM4
A100X
H100 PCIE
H100 SXM
RTX 4090
RTX 4080
L40
RTX 6000Ada
RTX 4080 S
RTX 4070
A40
A10
RTX A6000
RTX A5000
RTX A4000
RTX 3090
RTX 5000Ada
RTX 3090 Ti
RTX 3080 Ti
RTX 3080
RTX 3070
RTX 3060
Q RTX 8000
Tesla V100