Special call: GPU Testing and Benchmarking

Special call: GPU Testing and Benchmarking

! Update: the application deadline under this special call has been extended to 29 June 2023.

Computational resources to be distributed in NODE HOURS:

LUMI-G: 120,000 node hours

Maximum allocation per project: 8,000 node hours

Technical parameters of the GPU accelerated part of the LUMI supercomputer: The LUMI-G compute nodes are equipped with four AMD MI250X GPUs based on the 2nd Gen AMD CDNA architecture. A MI250x GPU is a multi-chip module (MCM) with two GPU dies named AMD Graphics Compute Die (GCD). Each of these Dies features 110 compute units (CU) and has access to a 64 GB slice of HBM memory for a total of 220 CUs and 128 GB total memory per MI250x module.

2-page application deadline: 
8–29 June 2023 

Eligible applicants: 
An employee of a Czech research organisation

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

The application form:

• DOCX format
• ODT format

Please submit the form via the web portal.

The period to use computational resources obtained:
The allocated computational resources can be used from 21 July 2023.

Call type: Open

Evaluation of applications:
The application will be evaluated by HPC experts from IT4Innovations only in terms of technical readiness to use the GPU-accelerated part of the LUMI supercomputer and the efficiency of the required number of node hours. The applicant will be notified of the outcome of the competition on 19 July 2023. Notification will be made electronically to the e-mail address provided in the application.

Mandatory acknowledgment 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 USERS' DUTIES.