Change machine type and configure GPUs of a Vertex AI Workbench instance

This page shows you how to change the machine type and GPU configuration of a Vertex AI Workbench instance.

Overview

The machine type determines some specifications of your Vertex AI Workbench instance, such as the amount of memory, virtual cores, and persistent disk limits. Changing the machine type can improve performance or help avoid errors caused by high memory utilization.

GPUs provide hardware acceleration that can improve the performance of your notebook code. You may want to add or increase the number of GPUs for greater performance, or, to reduce the cost of running your Vertex AI Workbench instance, you may want to remove GPUs.

Before you change a machine type or GPU configuration

Consider the following before you make any changes to your Vertex AI Workbench instance's machine types or GPU configuration.

Billing implications

Each machine type and GPU configuration is billed at a different rate. Make sure you understand the pricing implications of making a change. For example, an e2-highmem-2 machine type costs more than an e2-standard-2 machine type.

Changing between G2 and non-G2 machine types not supported

Changing between a G2 machine type and a non-G2 machine type isn't supported. If your instance has a G2 machine type, you can only change the machine type to a different G2 machine type. If your instance has a non-G2 machine type, you can change the machine type to a different non-G2 machine type, such as an E2 or N2 machine type.

Moving to a smaller machine type

If you move from a machine type with more resources to a machine type with fewer resources, such as moving from an e2-standard-8 machine type to an e2-standard-2, you might run into hardware resource issues or performance limitations because smaller machine types are less powerful than larger machine types.

Make a backup

It's good practice to make regular backups of your instance's persistent disk data using snapshots. Consider taking a snapshot of your persistent disk data before you change the machine type. If you want to make sure the new machine type is able to support the data on the existing instance, you can take a persistent disk snapshot and use it to start a second instance with the new machine type to confirm that the instance starts up successfully.

Change the machine type and configure GPUs

To change the machine type or configure the GPUs on a Vertex AI Workbench instance, complete the following steps.

  1. In the Google Cloud console, go to the Instances page.

    Go to Instances

  2. In the Notebook name column, click the name of the instance that you want to modify.

    The Notebook details page opens.

  3. Click the Hardware tab.

  4. In the Modify hardware configuration section, select the Machine type that you want to use.

  5. If there are GPUs available for your instance's combination of zone, environment, and machine type, you can configure the GPUs. In the GPUs section, select the GPU type and Number of GPUs that you want to use. Learn more about GPU regions and zone availability.

  6. If you haven't already installed the required GPU drivers on your instance, select Install NVIDIA GPU driver automatically for me to install the drivers automatically on next startup.

  7. After Vertex AI Workbench has finished updating the machine type and GPU configuration, you can start your Vertex AI Workbench instance.

What's next