0

My GPU quota has been increased on GCP and it shows as available, but for some reason cannot access it.

I tried various regions, also deleted the previous instance with the smaller quota within the same project. However, anytime I try the new quota it throws an error of exceeding the quota. I was able to recreate an instance with the previous, smaller quota.

Here is the code: gcloud compute instances create $INSTANCE_NAME --zone=$ZONE --image-family=$IMAGE_FAMILY --image-project=my-project --maintenance-policy=TERMINATE --accelerator="type=nvidia-tesla-p4,count=2” --machine-type=$INSTANCE_TYPE --boot-disk-size=200GB --metadata="install-nvidia-driver=True”

The quota can be also be requested via GCP Console as described in other post. This is not a question on how to apply for a quota increase, but rather why after getting Global GPU quota one cannot access it in any of the regions.

corio
  • 25
  • 6
  • 1
    Possible duplicate of [How to request GPU quota increase in Google Cloud](https://stackoverflow.com/questions/45227064/how-to-request-gpu-quota-increase-in-google-cloud) – Martin Zeitler Apr 06 '19 at 04:12

1 Answers1

0

Google Support explained that "NVIDIA P4 is one of the models of GPUS which you have to file a separate request aside from GPUS_ALL_REGIONS.", so simply had to file an additional GPU request specific to the region.

corio
  • 25
  • 6