unable to increase the NVIDIA_A100_GPUS quota

[MIRSK Development]

I get the following error when i try to create a us-central1-a VM with a2-highgpu-8g gpu's:

Quota 'NVIDIA_A100_GPUS' exceeded. Limit: 0.0 in region us-central1.

Going to the Quota page i can see that the limit for all regional quotas for compute.googleapis.com/nvidia_a100_gpus are unlimited and they are disabled (grayed out), so i cannot increase the quota for any of them.

This is not a problem when i increase the CPU quotas, but then again all our VM's except this one is located in europe-west1-b. Problem is that i think it is only us-central1-a that has the GPU option.

What should i do to overcome this problem?

[Fady (Google Cloud Platform)]

Generally to verify GPU availability, you may use this table and filter it by location and machine type.  In your case, the error indicates you do not have A100 GPU quota in region us-central1. This is different from Global Project quota (not all resources) . Now I believe that the boxes that are grayed out are for zones and not regions, and I assume that GPUs are regional resources and not zonal. A way to work around it is to filter location:us-central-1. This should only display the region and the box should not be grayed . Hence, you should be able to request quota in that region. I hope this helps. 

[MIRSK Development]

hi Fady

Thank you for you help. Your filter worked. I found the one i needed and have now requested an increase in GPU quota.