LUMI Sweden Allocations
Calls for LUMI Sweden allocations are issued annually by SNIC. The allocation period start on 1 January for a 12-month period. The LUMI Sweden Fall 2022 call was open through 14 October 2022 at 15:00 CEST.
For the call LUMI Sweden Fall 2022, the procedure will be similar to SNIC Large where applicable. Please refer to the requirements for Large. In addition to the eligibility requirements for SNIC Large, experience from SNIC Large, PRACE, or similar is required in order to be eligible for allocations in this round. The calls for LUMI Sweden rounds are aligned in time with SNIC Large calls. LUMI is well suited for parallel compute jobs that cannot easily be run within the existing Swedish HPC infrastructure.
It is allowed to apply for SNIC Large and LUMI Sweden simultaneously, and allocations in both rounds may be granted. You are strongly encouraged to also apply for the JU part of EuroHPC.
For more information on the resource, please refer to https://www.lumi-supercomputer.eu/may-we-introduce-lumi/
LUMI consists of several partitions. This call makes available the GPU partition, the CPU partition, LUMI-C, and the storage partitions LUMI-F, LUMI-P, and LUMI-O. The table below shows the Swedish share of the resource that is available in the call LUMI Sweden Fall 2022 (period through 31 December 2023). CPU and GPU compute capacity is applied for using core hours and GPU hours respectively for the entire call. Your are encouraged to distribute your resource consumption as evenly as possible over the project lifetime since it cannot be guaranteed that there is capacity available at the end of the project duration.
Please notice that storage has to be applied for within the call, if needed. Storage on LUMI is billed on a running cost basis and you need to apply for a total number of TB x hours, not the maximum disk quota. Storage is applied for using TB hours. Flash storage, LUMI-F, is accounted at ten times the TB-hour rate, i.e. use of 1TB of Flash storage for one hour costs 10 TB hours. Lustre storage, LUMI-P, is accounted at the TB-hour rate. CEPH storage, LUMI-C, is accounted at ½ the TB-hour rate, i.e. use of 1TB of CEPH storage for one hour costs 0.5 TB hours. You may combine the three types of storage at your own discretion, as long as you stay within the amount granted to your project and do not exceed the quota for any of the storage types at any time. You may need to contact the LUMI help desk at the start of the project to raise your maximum disk quota if you have large storage allocations.
- LUMI Sweden Fall 2022: The total number of core hours/GPU hours/TB hours available for this round. N.B. for the entire round, not per month!
- Minimum: Represents the minimum allocations in core hours/GPU hours. An application for a LUMI Sweden allocation must include a request at or above this limit for at least one measure.
|Resource||LUMI Sweden Spring 2022||Minimum|
|LUMI-C||58 000 000||4 000 000|
|LUMI-G||3 000 000||200 000|
|LUMI Storage||29 000 000||no limit|
4 000 000 core hours corresponds to 333 000 core hours per month if spread out evenly during the allocation period.
How to apply for LUMI Sweden
Please begin by reading the general information on how to apply for allocations on SNIC resources.
In addition to the application form, the application must include the following appendices submitted electronically through the upload facility in the online proposal submission system.
If you have already applied for a SNIC Large project, you may want to reuse the documents from that application.
1. CV of the PI - principal investigator (Maximum 2 pages)
2. Publication list of the principal investigator
3. SNIC Project Description (Maximum 5 pages)
Project description for LUMI Sweden
The project description shall not exceed five A4 pages (6½ if you include a Data Management Plan) set in 11pt or larger with a line spacing of at least 1.15 or similar. The disposition should follow the major topics listed here.
- Overview (approx. ½ page)
Provide an abstract of the proposed research, computations, and planned analysis.
- Research group and management (approx. ½ page)
Describe the complete research group that will be working on the proposed project. For each member, please specify at least the name of the member, HPC experience, position, and roles within the project. Indicate how the allocated time will be managed within the group.
- Implementation (approx. 1½ pages)
Describe how your applications can make efficient use of the requested resources. Give numbers and/or indicate measures of scalability and performance (for example a graph of execution time versus the number of cores). Include records of usage if you have been granted time for this or a similar project in previous allocation rounds. If the requested time differs significantly from earlier allocations and/or used time, please explain the difference.
- Data management plan (approx. 1½ pages)
Describe how your applications can make efficient use of the requested resources. Motivate and provide explicit numbers related to size, generation rate, number of files et cetera. Describe how data will be generated (codes, data type et cetera), and transferred into and out of the system. Describe how data will be managed on the system, for example using ephemeral/permanent storage. Describe how your project can efficiently balance the use of high-performance storage and lower-cost storage, and how your dataflow will be set up to accommodate for that. If applicable, describe how data will be made publicly available following the end of the project.
- Scientific challenges (approx. 2 pages)
Describe how the proposed project relates to the state of the art of the scientific area, has the potential to increase the knowledge within its scientific area, and how the proposed project uses adequate computational methods to address scientific questions. Include a justification for the need for large-scale data/HPC.
References (approx. ½ page)
Include a list of references for the project description.