Create a GPU VM

GPU Virtual machines are Linux-based VMs running on virtualized hardware with high-end GPUs. Each VM created serves as a new virtual server usable standalone or within a larger cloud infrastructure.

Step 1: Open Instance Management

Navigate via the side menu to Compute Engine → Instance Management, then select Create Instance.

Step 2: Configure the Instance

1. Instance Name: Provide a unique name for the GPU VM.
2. Instance Type: Choose the instance type. GPU is optimized for high-performance computing, machine learning, and other intensive tasks. Supported GPUs include NVIDIA H100 SXM5 and NVIDIA H200 SXM.
3. Disk type: Only one disk type selectable during creation: Ephemeral Disk (NVMe) or Persistent Disk (Block storage)
4. Image: Use either the default Ubuntu base image or a custom image.
   • OS: Currently supports UBUNTU 22.04 GPU
   • Custom image: Upload under Custom Images; it then appears in the selection list. (File type is QCOW)
5. Resource type: Each GPU VM has different configurations for vCPU, RAM, and number of attached GPUs. Choose the configuration that fits your needs.

Step 3: Configure the Storage disk

• Storage Policy: Specifies the storage type for the GPU VM.

  • GPU VMs with Ephemeral Disk (NVMe) support only NVMe-SSD.
  • GPU VMs with Persistent Disk support only Premium SSD, offering IOPS between 3,000 and 10,000 depending on service quota.

  Size:

  • Ephemeral Disk (NVMe): Fixed capacity per GPU instance, depending on GPU count in the selected Resource type.
  • Persistent Disk: Scalable starting from 100GB.

Step 4: Configure the Network Settings

• Subnet: Select the appropriate subnet for internal and external connectivity.
• Advanced network:
  • Private IP: Enter manually or let the system auto-assign based on the selected subnet.
  • Floating IP: For Ephemeral disk NVMe, floating IP is configured only after the VM is successfully created.
  • Security group: Assign a security group to manage inbound/outbound traffic.

Step 5: Set Authentication Method

Choose one of the following:

• SSH Key: The system automatically uses your latest SSH key (changeable if needed).
• Password: Set a password and store it securely for console access.

Step 6: Advanced Settings

Backup Job

Only supports GPU VMs using Block storage - Persistent disk. You can schedule automatic backups with defined frequency and timing.

Backup Options:

• Daily Full Backup – Full backup every day.
• Daily Incremental, Weekly Active Full – Daily incremental with weekly full.
• Daily Incremental, Monthly Active Full – Daily incremental with monthly full.

Backup Time: Set the specific time for backup execution.

Tags

Assign existing tags to manage and categorize resources.

User Data (Cloud-init Script)

The User Data field allows adding cloud-init scripts. When the VM starts, cloud-init reads metadata and auto-configures the system including users, SSH keys, and network settings.

Sample Cloud-init Script: Creates user "testcloudinit" with password "Abc123" and user "testcloudinit2" with password "P@ssw0rd!".

users:
- name: testcloudinit
  sudo: ALL=(ALL) NOPASSWD:ALL
  lock_passwd: false
  shell: /bin/bash
  passwd: $6$rounds=4096$V6anciWl30$xKbcljqks1gUkMiM80pyKzhvyhn7U1n.jXcGCUfkUlX.rnllUWKUrmDEzekhhhP8aERSylRuC7gfDhJ32Xv0A1
- name: testcloudinit2
  groups: sudo
  lock_passwd: false
  shell: /bin/bash
  plain_text_passwd: P@ssw0rd!
- hostname: testcloudinit

Step 7: Billing Plans

When creating a GPU VM with Persistent Disk, choose from two billing plans:

Billing plans	Description
Reserved	GPU availability is guaranteed. Charges apply even when the VM is powered off.
PAYG	When powered off, the GPU is automatically detached — no GPU charges while stopped. Only storage and floating IP costs continue. GPU availability is not guaranteed when powering back on.

You can update the billing plan after creation as long as the VM is in the Running state. Changes take effect immediately and reflect in your bill.

Step 8: Create the Instance

Click Create Instance to deploy and start the GPU VM.

Once created successfully, view details in the Instance Management dashboard.