Nuvolos Compute Units (NCUs)

The Nuvolos Compute Unit (NCU) is the computational unit for regular (non-HPC) applications run by users, and is the main unit of account for Nuvolos subscriptions.

In hardware terms, one NCU consists of one virtual CPU (vCPU), 4 GB of RAM, and adequate temporary storage to run a typical Application.

A vCPU (virtual CPU) is a share of a physical processor core allocated to your application by the cloud infrastructure. It behaves like a dedicated CPU core for scheduling and running your code, but multiple vCPUs from different users may share the same physical hardware. A 4 CU application, for example, can use up to 4 vCPUs - roughly equivalent to a quad-core laptop processor.

CU consumption of applications

Every time a user starts an application, a corresponding number of CUs are checked out from the service. The following default values apply:

• Applications in teaching spaces consume a single CU.

• Applications in research and dataset spaces consume 4 CUs.

Space administrators can change the CU consumption of applications. In research spaces, even Instance Editors can change the CU settings of applications to allow for seamless collaboration. To learn how to resize an application, see Changing application resources.

Application sizes can be changed at any time as long as the project is active. Running applications must be restarted for the new size to take effect. Beyond the included sizes, credit-based sizes with GPUs and larger configurations are available - see Application resources for the full details.

CU limits of a resource pool

Each resource pool has a certain number of CUs available for consecutive use over the subscription period. To review your current utilisation, see CU limits and capacities.

When you reach your CU limit, applications continue to run without interruption as long as you maintain a positive credit balance. For options when approaching or exceeding capacity, see NCU limits and capacities.