Types of allocations
Rules for allocating allocations
Submitted applications are forwarded to the user support team, which, with the involvement of scientists from various fields, evaluates them in terms of formal, technical, and scientific aspects.
The assessment will take into account:
- scientific purpose of the research being carried out,
- the validity of the use of large-scale calculations, including massively parallel calculations,
- preparation the applicant and his team to realization of the project,
- the technical feasibility of the project on a given supercomputer, including the availability of tools and the amount of the available resources,
- in the case of research teams that have already performed calculations in ICM, one of the elements the results obtained from previous projectes will also be assessed.
- projects requiring the use of large-scale calculations with breakthrough importance in their fields of science are particularly promoted.
To ensure the optimal use of the infrastructure, ICM reserves the right to allocate slightly fewer computational resources than requested or to propose an alternative available computing system.
Publications resulting from the use of ICM’s computational resources should include an appropriate acknowledgment in the form of:
This research was carried out with the support of the Interdisciplinary Centre for Mathematical and Computational Modelling at the University of Warsaw (ICM UW) under computational allocation no. Gxx-xx.
Alternatively, an equivalent wording may be used, adapted to the style and character of the text.
The duration of the allocation
By default, access to a supercomputer is granted for 1 per year. If the allocated allocation is not used, it is possible to extend it for another year.
Deadlines for submitting applications
Applications are accepted continuously.
Types of Allocations
The allocation type depends on the amount of requested CPUh and a supercomputer (Okeanos, Topola, Rysy) on which the calculations are to be performed.
| Cluster | CPUh | GPUh | Assessment | Remarks |
|---|---|---|---|---|
| Topola | 1 000 000 | - | technical | |
| Rysy | individual | individual | technical | |
| Okeanos | < 20 000 | - | none | for tests |
| Okeanos | < 2 000 000 | - | technical | |
| Okeanos | > 2 000 000 | - | technical + scientific | e.g. grants awarded by NCN, FNP, etc. |
Detailed description is provided below.
Topola
On the Topola supercomputer, only one type of allocation is awarded, amounting to 1 million core hours.
Rysy
Due to the relatively small size and specificity of the Rysy supercomputer (GPU cluster), allocations are negotiated individually. This supercomputer is an experimental machine, therefore no dedicated software is installed in the operating system. The main usage is focused on fundamental research related to machine learning, with low computational complexity. The application should justify the need to use the GPU in the calculations.
Okeanos
The resources of the Okeanos supercomputer are made available primarily to scientists, whose research work requires the use of massively parallel calculations.
Great Challenges
These allocations are intended for teams requesting computing time in excess of 2 million core hours. Due to the amount of resources, both the requests for allocations and annoual reports are rigorously assessed.
Standard allocation
Standard allocations are granted to teams requesting computation time no more than 2 million core hours.
Test allocation
Allocations are granted to teams requesting low computing resources (~20,000 core hours) needed to test software or computational models. The main purpose of type test allocations is to collect information necessary to apply for a "Standard" or "Great Challenges" allocations. Applications will only be assessed in formal and technical terms. Computational resources for test allocations will be awarded for 6 months.