> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.umnai.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.umnai.com/_mcp/server.

# Processing compute configurations

A processing compute configuration defines the machine used by a data-processing job. Ingestion and onboarding jobs reuse it while setting their own inputs and runtime limits. Create it before submitting either kind of job.

## Before you begin

Identify the size and processing needs of your source data. The machine selector shows the processing machines available in your environment, with CPU and memory specifications.

## Create a processing compute configuration

1. Open **Data Lab > Compute Configurations** and select **Create**.
2. Enter a **Configuration name** that identifies the intended processing setup.
3. Open **Machine type** and search the available options. Compare the displayed CPU and memory specifications, then select a machine.
4. Select **Create**. Open the saved configuration and check its name, ID, **Compute type** and **Machine type**.

| Field              | What to supply                                        |
| ------------------ | ----------------------------------------------------- |
| Configuration name | A non-empty name for this reusable machine selection. |
| Machine type       | A processing machine from the supplied list.          |

The compute type for this form is **PROCESSING**. Training uses a separate [training compute configuration](/ui-guides/models/training-compute).

## Review and reuse a configuration

In **Overview**, check the machine and audit information. Use **Details** for the underlying record and identifiers. Select the saved configuration in the compute field of an ingestion or onboarding job.

Set the runtime on each job: ingestion uses seconds, while onboarding uses minutes. Creating a compute configuration saves a machine selection; it does not run a processing job.

## Rename or replace a configuration

Open the saved record and select **Edit** to change **Configuration name**, then save. Compute type and machine type are read-only on the edit form. Create another configuration when you need a different machine, and select it for the next job.

Use **Delete** on a configuration only after checking that it is no longer needed. Review the confirmation before removing the record.

## Next steps

Prepare a [data schema](/ui-guides/data/schemas), then run [ingestion](/ui-guides/data/ingestion-jobs).