LUMI Access

LUMI, one of the pan-European pre-exascale supercomputers located in Finland, is governed by a consortium of 10 countries. The consortium countries manage half of the resources and half of them are governed by EuroHPC Joint Undertaking in European-wide calls and managed by PRACE, similar to PRACE calls. Resources managed by Finland are described below.

All resources are awarded in three units: GPU hours, CPU core-hours and storage-hours, not in billing units as for CSC’s servers (Mahti, Puhti, cPouta, ePouta, etc.). LUMI consists of a CPU partition (LUMI-C), GPU partition (LUMI-G, with most of LUMI performance) and other services to be described later (object storage LUMI-O etc).

LUMI-G will be available only in the second phase of LUMI installation in the spring 2022.

Finnish LUMI access modes

Extreme Scale Access

  • largest projects using especially GPUs (LUMI-G)
  • 50% of Finnish LUMI GPU resources, 25% of Finnish CPU resources
  • duration fixed 12 months
  • 0.5–2 million GPU-h, max 16 million CPU core-h
  • calls for projects once a year
  • scientific evaluation by an external panel

Regular Access

  • middle size projects more confined to CPUs (LUMI-C)
  • duration fixed 12 months
  • max 8 million core-h, 0.2 million GPU-h, 570 000 TiB-h
  • application in MyCSC (to be announced)
  • project and award decisions every three weeks in CSC Resource allocation group meetings

Benchmark Access

  • to show readiness for Extreme Scale Access in LUMI-G (cf. PRACE preparatory access)
  • duration fixed 3 months
  • max 0.5 million core-h, 0.05 million GPU-h, 22 000 TiB-h
  • application in MyCSC (to be announced)
  • project and award decisions in a few days by CSC Resource management

Development Access

  • minor projects to develop software for LUMI
  • duration 1 year, extendable for other years
  • max 1 million core-h, 0.1 million GPUh, 90 000 TiB-h
  • application by email to Servicedesk including: title, project description, science area, program codes, methods, and development objectives
  • project award decisions every three weeks in CSC Resource allocation group meetings

Adaptation to LUMI-G in Extreme Scale Access

The codes in the proposed Extreme Scale Access projects need to be compatible with the LUMI-G architecture and its next generation AMD CPUs and, especially, GPUs.

The proposal must justify the technical feasibility of the intended applications: explain why and how the code will be able to be run AMD GPUs on LUMI.  

Codes already GPU enabled may be ported to LUMI-G. For example, codes using CUDA can typically be converted to use HIP with modest effort and good performance. GPU offloading with OpenMP or OpenACC is also supported to certain extent.

For more information, please visit:

LUMI EAP (Early Access Platform) with MI100 GPUs will be available in October for code development and testing.

We encourage the applicants to contact servicedesk@csc.fi if they need any further information regarding the compatibility of their codes or support in LUMI porting.