Problem deploying A100 40GB on designed region.

Hi, I’m trying to deploy a VMI with a A100 40gb in europe-west4-a where the docu says it’s available. I’m using the NVIDIA GPU-Optimized VMI from the marketplace. I want to know if this GPU needs an allocation quota (I think only the 80GB one needs it) and why It’s not available.

This might happen only in marketplace images, the GPU is avalaible deploying the vmi with the default settings.

Thanks in advance!

Hi @idominguez ,

Welcome to Google Cloud Community!

When deploying a Virtual Machine Instance (VMI) with an NVIDIA A100 40GB GPU in a specific region, such as europe-west4-a, you usually need to request an allocation quota in Google Cloud when you’re working with GPUs, especially with high-demand like the A100. For this specific GPU, only these regions/zones are available.

To check your current quota and usage, you should verify your GPU quotas in the Google Cloud Console. Navigate to the “IAM & Admin” section, select “Quotas & System Limits” and on the filter part, type “GPU”.

In addition, you can request a higher quota value, by following these steps:

  1. Go to the Quotas & System Limits page. Use the Filter search box to search for your quota. (name: NVIDIA A100 GPUs and zone: europe-west4-a)
  2. Select the checkbox to the left of your quota.
  3. Click create Edit. The Quota changes form displays.
  4. In the Quota changes form, enter a “new quota value or choose unlimited” for your project in the New value field. Once done, click confirm.
  5. If you choose Unlimited, click Done. and then click Submit request.

Please note that, both the NVIDIA A100 40GB and A100 80GB GPUs typically require GPU quotas. However, it’s important to note that the 80GB variant might have more strict quota requirements or be subject to specific allocation policies compared to the 40GB variant.

You can visit Compute Engine quota and limits and Allocation quotas availability for your reference.

I hope the above information is helpful.