The FENIX infrastructure is currently able to provide researchers with access to resources as listed in the table below:
| Component |
Site (Country) |
Minimum Request |
Total Resources offered | Unit |
|---|---|---|---|---|
| Cloud Computing Services | ||||
| Cluster OpenStack |
CEA (FR) | 1 VM | 320 vCPUs (no vGPUs) with a total of 640 GB RAM |
vCPUs |
| JSC Cloud | JSC (DE) | 1 VM | 1500 vCPUs, 3000 GB Memory | vCPUs, Memory |
| Archival data repositories | ||||
| Object Store S3* | CEA (FR) | 1 VM | 500 ** | TB |
Further details on the components are provided below:
Components from CEA (France)
- Archival: Object store accessible through the S3 protocol.
- Cluster OpenStack
Accessing from OpenStack cluster to the object store requires a TGCC user account.
Components from JSC (Germany)
- JSC Cloud: A subset of the hardware listed at the official documentation is available for FENIX. A specific amount of the mentioned AMD EPYC Rome 7742 64-core CPU nodes that feature 256 GB DDR4 memory. Nodes with an additional Nvidia Volta V100 GPU with 16 GB high-bandwidth memory are available.
- Research groups that have already allocated resources on JSC Cloud are eligible to apply. However, such projects will be given lower priority in the evaluation process than projects that have not previously used resources from this cloud infrastructure.
In particular, projects with active allocations will
1. Be assessed to justify the need for additional resources, including why existing allocations cannot meet their needs.
2. Have their application ranked lower unless they demonstrate significant new scientific or technical contributions.
This measure will ensure equitable access to resources and encourage the participation of new users and projects.
** Not possible to access the S3 storage from the OpenStack cluster, but the storage can be used for back-up using the FENIX FTS service
