Products

For each selected grant giver, the application form displays a list of available resources categorized by the provider.

Details such as remaining resources and allocation time can be viewed by navigating to Allocations and Usage under the Project section of the navigation menu.

Note

The updated costs of each product are listed here.

drawing DeiC Interactive HPC (SDU/K8s)

Storage

The storage service includes one product:

  • Weka File System: storage

When applying for storage resources, users need to specify the total amount of resources for these products in GB. Data can be stored for a duration at least as long as the allocation grant period.

Note

If the current allocation expires, data will not be removed from UCloud. However, users will be unable to add more data or initiate new jobs using the existing data.

Compute

Users can select from different products offered by the UCloud compute service. Each product is subdivided into various machine types:

  • SDU/K8s CPU nodes: cpu-amd-zen5

    A CPU node consists of 2x AMD EPYC 9635 CPU@2.25 Ghz, 128 vCPUs, and 768 GB of DDR5-6000 memory.

  • SDU/K8s GPU nodes: gpu-nvidia-b200

    A GPU node consists of 2x AMD EPYC 9655 CPU@2.6 Ghz, 384 vCPUs, and 2304 GB of DDR5-6400 memory; 8x NVIDIA Blackwell B200-SMX6, 192 GB.

  • SDU/K8s Micro Instance GPUs (MIG): gpu-nvidia-b200-*-mig

    One Micro Instance GPU (MIG) consists of 1/7 of a NVIDIA B200 GPU card (see hardware specification above). Users can request between 1 and 4 MIGs for a job.

Note

MIGs are accounted with gpu-nvidia-b200 resources and one MIG costs 1/7 GPU-hour/hour.

Public IPs

UCloud apps can be deployed with a static public IP address, facilitating access by external third-party applications.

Note

This feature is only available on a limited set of applications. Allocated IPs can be released if they remain unused for an extended period.

Licenses

Depending on the grant giver, users can also apply for specific software licenses within a project. Software licenses are listed in the Resource allocations panel of the project dashboard, similar to other products.

Note

Access to the license server is controlled by the project admins, as described here.

drawing SDU-Odense

Note

This provider is only available to users affiliated with the University of Southern Denmark.

Storage

The storage service includes one product:

  • Weka File System: storage

When applying for storage resources, users need to specify the total amount of resources for these products in GB. Data can be stored for a duration at least as long as the allocation grant period.

Note

If the current allocation expires, data will not be removed from UCloud. However, users will be unable to add more data or initiate new jobs using the existing data.

Compute

Users can select from different products offered by the UCloud compute service. Each product is subdivided into various machine types:

  • SDU/K8s CPU nodes: cpu-amd-zen4

    A standard node consists of 2x AMD EPYC 9534 CPU@2.45Ghz, 128 vCPUs, and 768 GB of DDR5-4800 memory.

  • SDU/K8s GPU nodes: gpu-nvidia-h100

    A GPU node consists of 2x AMD EPYC 9454 CPU@2.75 GHz, 96 vCPUs, and 768 GB DDR5-4800 of memory; 4x NVIDIA H100 SXM5 GPU accelerators 80 GB.