Skip to main content

Create your first compute environment

Quick start: create a volume, create your first compute resource (for example a Jupyter notebook with PyTorch on a GPU) and list your volumes and compute resources.

Creating a volume​

cgc volume create <name> --size <size> -sc <storage_class>

  • -s - size in GiB of your volume
  • -sc, --storage-class - storage class of your volume

For the created volume to be accessible, it needs to be mounted to a resource.

Output

cgc volume create getting-started -s 10 -sc <storage_class>
Volume getting-started of size 10.0 GB on SSD created. Volume is ReadWriteMany.

List of existing volumes​

List all created volumes:

cgc volume list

example:

cgc volume list
name             used    size    type           disks type    mounted to
---------------  ------  ------  -------------  ------------  ------------
getting-started  7.67M   10Gi    ReadWriteMany  SSD

Compute create​

To create your first compute resource, you need to specify the name and type of entity for the compute resource.

cgc compute create <entity> --name --gpu -gpu-type --cpu --memory --volume

where

  • entity - i.e. nvidia-pytorch
  • -n, --name - name of the compute resource
  • -g, --gpu - quantity of attached GPUs
  • -gt, --gpu-type - type of attached GPU, named as in cgc status (NVIDIA-RTX-A5000 | NVIDIA-A100-SXM4-80GB | NVIDIA-B200) default = NVIDIA-RTX-A5000
  • -c, --cpu - cpu core count
  • -m, --memory - amount of attached RAM in GiB
  • -v, --volume - volume to mount name of volume

example:

cgc compute create -n getting-jupyter -c 2 -m 4 -g 1 -gt NVIDIA-RTX-A5000 -v getting-started nvidia-pytorch

nvidia-pytorch app getting-jupyter creation started!
 Volumes to mount: getting-started
Container will be Ready after volumes can be see mounted.
Will be accessible at: https://getting-jupyter.namespace.cgc-waw-01.comtegra.cloud
App token: f660c69776c647eba6d895c312388a0c
To monitor the startup status use list command
Environment variables: JUPYTER_PORT=8888, JUPYTER_TOKEN=f660c69776c647eba6d895c312388a0c

Output provides an accessible URL and generated token. Token can be changed at first login.

List of existing compute resources​

If you want to see the list of your existing compute resources, you can use the list command.

cgc compute list

example:

cgc compute list
name             type            status    volumes_mounted      cpu  ram      count    gpu-count  gpu-label         url
---------------  --------------  --------  -----------------  -----  -----  -------  -----------  ----------------  -----------------------------------------------------
getting-jupyter  nvidia-pytorch  Running   getting-started        2  4Gi          1            1  nvidia-rtx-a5000  https://getting-jupyter.namespace.cgc-waw-01.comtegra.cloud

To see Jupyter token add -d flag for more details.