Personal Compute is a Databricks-managed default cluster policy available on all Databricks workspaces. The policy allows users to easily create single-machine compute resources for their individual use so they can start running workloads immediately, minimizing compute management overhead.
Personal Compute resources are all-purpose clusters (priced according to all-purpose compute pricing) with the following properties:
Personal Compute resources are single-node clusters (that is, “clusters” with no workers and with Spark running in local mode).
They use the Single user cluster access mode and are Unity Catalog-compatible.
The runtime is set to use the latest version of the Databricks Runtime for Machine Learning.
Both standard instances and GPU-enabled instances are available
Auto-termination is set at 72 hours.
If you don’t see the Personal Compute policy as an option when you create a cluster, then you have not been given access to the policy. Please contact your administrator to request access to the Personal Compute policy or an appropriate equivalent policy.
You can create a Personal Compute cluster using either the Compute page or directly from a notebook.
You can create a Personal Compute resource from a notebook in only three steps:
At the top of your notebook, click the Connect button.
Select Create new resource…. This will open a dialog that defaults to the Personal Compute policy.
Once your Personal Compute resource is running, your notebook will connect to it automatically.
The Compute page includes a shortcut button for creating a Personal Compute resource:
Click Compute in the sidebar.
Click Create Personal Compute. This will open the cluster configuration dialog with the Personal Compute policy chosen.
Click Create Cluster.
Once the cluster starts, your Personal Compute resource will be available for your use.