Console

1. In the Google Cloud console, go to theCreate an instance page.

   Go to Create an instance

2. Specify aName for your VM. SeeResource naming convention.

3. Select a region and zone where GPUs are available. See the list of availableGPU zones.

4. In theMachine configuration section, select theGPUs machinefamily, and then do the following:

   1. In theGPU type list, select one of the GPU models supported onN1 machines.
   2. In theNumber of GPUs list, select the number of GPUs.

   3. If your GPU model supportsNVIDIA RTX Virtual Workstations (vWS) for graphics workloads, and you plan on runninggraphics-intensive workloads on this VM, selectEnable Virtual Workstation (NVIDIA GRID).

   4. In theMachine type list, select one of the preset N1 machinetypes. Alternatively, you can also specify custom machine type settings.

5. In theBoot disk section, clickChange. This opens theBoot disk configuration page.

6. On theBoot disk configuration page, do the following:

   1. On thePublic images tab, choose asupported Compute Engine imageorDeep Learning VM Images.
   2. Specify a boot disk size of at least 40 GB.
   3. To confirm your boot disk options, clickSelect.

7. Optional: In theVM provisioning model list, select aprovisioning model.

8. To create and start the VM, clickCreate.

gcloud

To create and start a VM use thegcloud compute instances createcommand with the following flags.

If your workload is fault-tolerant or can start at any time, then considerusing a different provisioning model to reduce your costs. To change yourprovisioning model, include the--provisioning-model=<var>PROVISIONING_MODEL</var> flag in the command.For more information about the available models, seeCompute Engine instances provisioning models.

gcloud compute instances createVM_NAME \    --machine-typeMACHINE_TYPE \    --zoneZONE \    --boot-disk-sizeDISK_SIZE \    --accelerator type=ACCELERATOR_TYPE,count=ACCELERATOR_COUNT \    [--imageIMAGE | --image-familyIMAGE_FAMILY] \    --image-projectIMAGE_PROJECT \    --maintenance-policy TERMINATE \    [--provisioning-model=PROVISIONING_MODEL]

Replace the following:

• VM_NAME: thenamefor the new VM.
• MACHINE_TYPE: themachine typethat you selected for your VM.
• ZONE: thezone for theVM. This zone must support theGPU type.
• DISK_SIZE: the size of your boot disk in GB.Specify a boot disk size of at least 40 GB.

• IMAGE orIMAGE_FAMILY thatsupports GPUs.Specify one of the following:

  • IMAGE: the required version of apublic image. For example,--image debian-10-buster-v20200309.
  • IMAGE_FAMILY: animage family. Thiscreates the VM from the most recent, non-deprecated OS image. Forexample, if you specify--image-family debian-10,Compute Engine creates a VMfrom the latest version of the OS image in the Debian 10 image family.

  You can also specify a custom image orDeep Learning VM Images.

• IMAGE_PROJECT: theCompute Engineimage projectthat the image family belongs to. If using a custom image orDeep Learning VM Images, specify the project that those images belong to.

• PROVISIONING_MODEL: the provisioning model forthe VM. Specify eitherSPOT orFLEX_START. Ifyou don't specify a provisioning model, the standard model is used. Thisflag is optional.

• ACCELERATOR_COUNT: the number of GPUs that you wantto add to your VM. SeeGPUs on Compute Enginefor a list of GPU limits based on the machine type of your VM.

• ACCELERATOR_TYPE: theGPU model that you want to use.If you plan on running graphics-intensive workloads on this VM, use oneof thevirtual workstation models.

  Choose one of the following values:

  • NVIDIA GPUs:

    • NVIDIA T4:nvidia-tesla-t4
    • NVIDIA P4:nvidia-tesla-p4
    • NVIDIA P100:nvidia-tesla-p100
    • NVIDIA V100:nvidia-tesla-v100

  • NVIDIA RTX Virtual Workstation (vWS) (formerly known as NVIDIA GRID):

    • NVIDIA T4 Virtual Workstation:nvidia-tesla-t4-vws
    • NVIDIA P4 Virtual Workstation:nvidia-tesla-p4-vws

    • NVIDIA P100 Virtual Workstation:nvidia-tesla-p100-vws

      For these virtual workstations, an NVIDIA RTX Virtual Workstation (vWS)license is automatically added to your instance.

Example

For example, you can use the followinggcloud command to start anUbuntu 22.04 VM with 1 NVIDIA T4 GPU and 2 vCPUs in theus-east1-d zone.

gcloud compute instances create gpu-instance-1 \    --machine-type n1-standard-2 \    --zone us-east1-d \    --boot-disk-size 40GB \    --accelerator type=nvidia-tesla-t4,count=1 \    --image-family ubuntu-2204-lts \    --image-project ubuntu-os-cloud \    --maintenance-policy TERMINATE

REST

Identify the GPU type that you want to add to your VM. Submit aGET request to list the GPU types that are available to your project in aspecific zone.

To create VMs at a discounted price, you can specify adifferent provisioning model by adding the"provisioningModel":"PROVISIONING_MODEL" field to thescheduling object in your request. For more information about the available models, seeCompute Engine instances provisioning models.

GET https://compute.googleapis.com/compute/v1/projects/PROJECT_ID/zones/ZONE/acceleratorTypes

Replace the following:

• PROJECT_ID: project ID.
• ZONE:zonefrom which you want to list the available GPU types.

Send a POST request to theinstances.insert method.Include theacceleratorType parameter to specify which GPU type you want to use, andinclude theacceleratorCount parameter to specify how many GPUs you wantto add. Also set theonHostMaintenance parameter toTERMINATE.

POST https://compute.googleapis.com/compute/v1/projects/PROJECT_ID/zones/ZONE/instances{  "machineType": "projects/PROJECT_ID/zones/ZONE/machineTypes/MACHINE_TYPE",  "disks":  [    {      "type": "PERSISTENT",      "initializeParams":      {        "diskSizeGb": "DISK_SIZE",        "sourceImage": "projects/IMAGE_PROJECT/global/images/family/IMAGE_FAMILY"      },      "boot": true    }  ],  "name": "VM_NAME",  "networkInterfaces":  [    {      "network": "projects/PROJECT_ID/global/networks/NETWORK"    }  ],  "guestAccelerators":  [    {      "acceleratorCount":ACCELERATOR_COUNT,      "acceleratorType": "projects/PROJECT_ID/zones/ZONE/acceleratorTypes/ACCELERATOR_TYPE"    }  ],  "scheduling":  {    ["automaticRestart": true],    "onHostMaintenance": "TERMINATE",    "provisioningModel": "PROVISIONING_MODEL"  },}

Replace the following:

• VM_NAME: thename of the VM.
• PROJECT_ID: your project ID.
• ZONE: thezone for theVM. This zone must support theGPU type.
• MACHINE_TYPE: themachine typethat you selected for the VM. SeeGPUs on Compute Engineto see what machine types are available based on your chosen GPUcount.

• IMAGE orIMAGE_FAMILY: specify one of the following:

  • IMAGE: the required version of a public image. For example,"sourceImage": "projects/debian-cloud/global/images/debian-10-buster-v20200309"
  • IMAGE_FAMILY: animage family. Thiscreates the VM from the most recent, non-deprecated OS image. Forexample, if you specify"sourceImage": "projects/debian-cloud/global/images/family/debian-10",Compute Engine creates a VMfrom the latest version of the OS image in the Debian 10 image family.

  You can also specify a custom image orDeep Learning VM Images.

• IMAGE_PROJECT: theCompute Engineimage projectthat the image family belongs to. If using a custom image orDeep Learning VM Images, specify the project that those images belong to.

• DISK_SIZE: the size of your boot disk in GB.Specify a boot disk size of at least 40 GB.

• NETWORK: the VPCnetwork that you want to usefor the VM. You can specifydefault to use your default network.

• ACCELERATOR_COUNT: the number of GPUs that youwant to add to your VM. SeeGPUs on Compute Enginefor a list of GPU limits based on the machine type of your VM.

• ACCELERATOR_TYPE: theGPU model that you want to use.If you plan on running graphics-intensive workloads on this VM, use oneof thevirtual workstation models.

  Choose one of the following values:

  • NVIDIA GPUs:

    • NVIDIA T4:nvidia-tesla-t4
    • NVIDIA P4:nvidia-tesla-p4
    • NVIDIA P100:nvidia-tesla-p100
    • NVIDIA V100:nvidia-tesla-v100

  • NVIDIA RTX Virtual Workstation (vWS) (formerly known as NVIDIA GRID):

    • NVIDIA T4 Virtual Workstation:nvidia-tesla-t4-vws
    • NVIDIA P4 Virtual Workstation:nvidia-tesla-p4-vws

    • NVIDIA P100 Virtual Workstation:nvidia-tesla-p100-vws

      For these virtual workstations, an NVIDIA RTX Virtual Workstation (vWS)license is automatically added to your instance.

• PROVISIONING_MODEL: the provisioning model for theVM. Specify eitherSPOT orFLEX_START. If youdon't specify a provisioning model, the standard model is used. Thisproperty is optional. For more information on provisioning models, seeCompute Engine instances provisioningmodels.