Institutions can apply for computational resources within Open Access Grant Competitions. The grant competition is announced three times a year (February, June, October) for research, scientific and educational organisation employees.

 

33Rd OPEN ACCESS GRANT COMPETITION OF IT4INNOVATIONS

 

Attention all applicants for KAROLINA GPU resources – KAROLINA GPU demand exceeds availability! Therefore, please apply for the GPU resources on the LUMI supercomputer.

Extensive resources on the GPU part of the LUMI supercomputer are available to employees of research organisations in the Czech Republic. Apply for resources for LUMI-G in the 33rd Open Access Grant Competition. LUMI-G offers excellent performance for artificial intelligence projects in PyTorch. For more information, visit https://docs.it4i.cz/lumi/.

Computational resources to be distributed (up to 25% of node hours of each resource per request): the total maximum possible allocations for each resource will be specified

  1. Barbora CPU: 460 000 node hours
  2. Barbora GPU: 20 000 node hours
  3. Barbora FAT: 2 600 node hours
  4. DGX-2: 1 200 node hours
  5. Karolina CPU: 950 000 node hours
  6. Karolina GPU: 70 000 node hours 
  7. Karolina FAT: 1 000 node hours
  8. LUMI-C: 150 000 node hours
  9. LUMI-G: 150 000 node hours

Technical parameters of the Barbora supercomputer:
192 compute nodes without accelerators, 8 compute nodes with 4x NVIDIA Tesla V100-SXM2 GPU accelerators, 1 FAT node for large-scale data analysis equipped with 6 TB of shared memory, 2 visualization nodes.

Technical parameters of the Karolina supercomputer:
720 compute nodes without accelerators, 72 compute nodes with 8x NVIDIA A100 GPU accelerators (40 GB HBM2), 1 FAT node for big data analytics with up to 24 TB shared memory, 2 visualization nodes, 36 nodes for cloud services.

Technical parameters of the NVIDIA DGX-2 system:
1 compute node with 2x Intel Xeon Platinum processors and 16x NVIDIA Tesla V100 GPU accelerators (32 GB HBM2).

Technical parameters of the LUMI supercomputer:
LUMI-G consist of 2560 nodes, each node with one 64 core AMD Trento CPU and four AMD MI250X GPUs.
LUMI-C is using CPU only nodes, featuring 64-core 3rd-generation AMD EPYC™ CPUs, and between 256 GB and 1024 GB of memory.

Application deadline:
3 October – 27 November 2024

Eligible applicants: Employee of a Czech research organisation

How to submit an application:
Applications will only be accepted in English.

The application form:

Please submit the form via the web portal.

Date of announcement of results:
24 January 2025

The period to use computational resources obtained:
The drawdown will be granted for a period of 12 months from the start of the drawdown, which is expected to start on 30 January 2025.
For multi-year projects, depending on the chosen project duration, the drawdown will be granted for 24 or 36 months from the start of the drawdown.

Call type: Open

Evaluation of applications:

Three areas will be evaluated: Scientific excellence, Computational readiness and Socioeconomic impact. The evaluation and scoring will be carried out by external reviewers, following the Open Access evaluation guidelines. Scientific excellence, Computational readiness and Socioeconomic impacts of the application will be evaluated. The allocation committee will take into account the connection of the project to other projects or grants: international (H2020, EuroHPC, ERC, …), national (GAČR, TAČR, …) or local (Student grant competition, …), to the extent and detail of requested resource estimation and to the expected applied outputs such as patents or contracts.
The weight of 50% will be assigned to the publications achieved as a result of previous projects at IT4I, registered at https://scs.it4i.cz/publications. On average 1 publication per project is expected. The latter will be only applied to projects older than 15 months. Applicants will be notified of the competition results within 35 working days of the conclusion of the competition. Notifications will be sent in electronic format to the e-mail address specified in the application.

Mandatory acknowledgement of the achieved R & D results:
This work was supported by the Ministry of Education, Youth and Sports of the Czech Republic through the e-INFRA CZ (ID:90254).

For the correct assignment of e-INFRA CZ to the R&D results when exporting them to RIV, see USER'S DUTIES. 

Date of announcement of the next open access competition:
February 2025

multi-year approach

You can also apply for projects in the form of "multi-year" open access, in which computing resources are provided for a period of 24 or 36 months. The purpose is to support long-term scientific grants.

More about the multi-year approach

Competition results