See more in link

Step 1: install nd setup gcloud SDK⌗

tutorial

Step 2: create vm⌗ 

export IMAGE_FAMILY="tf-latest-cu92"
export ZONE="us-west1-b"
export INSTANCE_NAME="my-new-instance"
export INSTANCE_TYPE="n1-standard-8"
export DISKSIZE="120GB"
gcloud compute instances create $INSTANCE_NAME \
        --zone=$ZONE \
        --image-family=$IMAGE_FAMILY \
        --image-project=deeplearning-platform-release \
        --maintenance-policy=TERMINATE \
        --accelerator="type=nvidia-tesla-v100,count=8" \
        --machine-type=$INSTANCE_TYPE \
        --boot-disk-size=$DISKSIZE \
        --metadata="install-nvidia-driver=True"

Step 3: control your vm⌗

  • list vm:
gcloud compute instances list
  • vm status
gcloud compute instances describe $INSTANCE_NAME
  • access vm
gcloud compute ssh $INSTANCE_NAME
  • stop vm
gcloud compute instances stop $INSTANCE_NAME
  • start vm
gcloud compute instances start $INSTANCE_NAME
  • delete vm
gcloud compute instances delete $INSTANCE_NAME