Competitive advantage through research ecosystems

A modern environment for research and RDI is a key factor for competitiveness and productivity. World-class supercomputers enable the efficient processing of large data sets and open new doors for research through the use of AI and quantum technology.

The data management and computing environments operated by CSC and the expert services associated with them mean that Finland punches above its weight as a cluster of expertize in computational science and enable state-of-the-art research in Finland and Europe.

International projects and infrastructures in which CSC is involved serve as one of the channels for the internationalization of Finnish research and higher education institutions.

CSC's Data Management and Computing Development programme (DL2021) ended in 2021. While this programme was in progress, the number of users of CSC's services increased by more than 40%. The number of research institute users has increased fivefold since 2017. Natural sciences are some of the largest users of CSC’s services out of the main fields of science, but technology, medicine and health sciences, humanities, agriculture and forestry and social sciences have also started using the services more during the programme.

The DL2021 programme also played an important role in piloting ways in which CSC's data management and computing service resources can be offered to companies. In these efforts, CSC worked together with Business Finland’s AI Business programme. The experiences gained through this programme, both in terms of technology and cooperation, where a contributing factor in EuroHPC’s decision to locate its pre-exascale supercomputer (LUMI) in Finland.

Steady growth in the use of CSC's research infrastructure services

The National Data Management and Computing Development programme (DL2021) supported the research and teaching activities of universities, universities of applied sciences and research institutes as well as national innovation activities, also securing the Finnish research community's international competitiveness in data-driven and computing-intensive fields of research. 

In a survey addressed to the users of CSC’s data management and computing services in 2021, the NPS of those who had used the DL2021 environment was 69.4. While this score can still be considered extremely good, it represents a slight drop compared to previous years. 

The final DL2021 investments comprised Mahti-AI, the AI partition of the supercomputer Mahti. The need for graphics processors (GPUs) has continued to grow strongly, and with the introduction of the Mahti-AI partition, it is now also possible to offer much needed GPU resources to Mahti users. Mahti-AI was opened for users in late April.

The final seminar of the DL2021 programme was held on 13 December 2021. CSC's self-assessment of the results and impacts of the programme can be found on the page CSC's self-assessment of the achievement of the DL2021 programme’s objectives.

Thanks to the National Data Management and Computing Development programme, research institutes have discovered CSC's services across a broad front, both as users and customers of new initiatives. In data management and computing services, the number of users from research institutes and universities increased significantly in 2021. The volume of CSC’s services used by teachers and students has also been clearly increasing, as has their use for international research.

The total number of users continued to grow in 2021. The table below describes the number of active users in CSC's research service package, including the IDA research data storage service. In this context, an active user refers to a user with a valid ID. Compared to other user groups, especially scientists at research institutes have increasingly discovered CSC’s services.

Numbers of CSC's computing and data management service users in higher education institutions and research institutes in 2018–2021.
  Universities Research institutes Universities of
applied sciences
Active users 2018 3281 232 95
Active users 2019 3671 384 233
Active users 2020 3633 422 449
Active users 2021 4088 512 467
Growth during 2021 12,5% 21,3% 4%

The fields of life sciences (459), computing and information sciences (396) and physics (230) accounted for the largest numbers of active research projects in CSC's server environments. These also were the largest fields when measured by the use of computer resources. 

The largest users of cPouta cloud computing resources were space science and astronomy (23%) as well as computing and information science (21%), while in ePouta, basic medical sciences (39%) and computing and information science (33%) accounted for the largest user groups.

Ten fields that were the largest resource users in 2021, development of use in 2018–2021 compared to other fields of science.

 

Shares of projects and resources used in 2021 by field of science. The fields Medical and health sciences, Engineering and technology, and Humanities contain all disciplines falling under the main category in question. ‘Other’ contains other fields of natural sciences besides those shown separately in the Figure, Social sciences and Agricultural sciences, and billing units that have not been allocated to any field of science (1% of all resources and 14% of projects). The total number of projects was 2,604, and the total volume of resources was 1,819 million billing units. Billing unit is the unit used by CSC to monitor the use of resources.

 

Small (15-35 active projects in 2021) disciplines where the number of active projects has at least doubled between 2018 and 2021.

 

The research and education services available through CSC are suitable for all fields of science and for supporting higher education teaching. Support is offered in all phases of the research process. CSC’s Resource Allocation Team grants customers national LUMI computing and storage resources on application. Rather than directly assessing the scientific quality of applications, CSC requires that the methods presented in the applications are technically viable and that the allocated resources are used to produce scientific publications. The resources must also be used responsibly for the needs of either research or teaching. The Resource Allocation Team monitors national science policy and observes the priorities set by Finland’s science administration.

Computing resources are primarily allocated to national research. Once a year, the allocated resources are reported to the CSC Board of Directors and the Ministry of Education and Culture. In 2021, the Resource Allocation Team processed 2,033 applications. In total, 1,742,259,522 billing units (BU) were applied for and 1,574,959,522 were granted. Allocations to Finnish users are granted out of Finland’s LUMI resources by CSC’s Resource Allocation Team. The most significant ones are the one-year Regular Access projects (Tier-1 level). The first Regular Access projects for LUMI-C were already granted in December 2021.

Applications for Grand Challenge research projects, which require significant computing or storage resources, and for international PRACE high-performance computing resources are assessed by the Scientific Customer Panel, which consists of Finnish research team leaders from a variety of scientific fields. Future Extreme Scale Access (Tier-0 level) projects included in Finland’s LUMI resources will be evaluated by a Nordic panel to be formed later.

The accessibility of the computing and data management environment was excellent in 2021 (Puhti 99.27%, 2020: 99.17% and Mahti 98.02%, 2020: 94.25%).

The performance enabled by CSC's computing servers has improved extremely rapidly in recent years. Year 2021 was Mahti's first full year in production, and its computing capacity corresponds to the total performance of all computing servers in 2020. Due to the rapid increase in the computing power, computer usage rates have fluctuated more than usual, or between 57% and 90% on a monthly basis. On the other hand, the total volume of computing increased by 64% compared to 2020.

Maximum peak performance and actual use between 2018 and 2021.

 

Phase 1 of LUMI supercomputer installed

Led by Finland, the LUMI consortium is responsible for hosting one of three world-class supercomputers organized by the European High Performance Computing Joint Undertaking (EuroHPC JU) as a joint procurement. When completed, the LUMI supercomputer will be one of the most well-known scientific computing instruments in the world during its service life.

Construction of the LUMI data center and system installationConstruction of the LUMI data center was completed in spring 2021 despite the global pandemic. Installation of the LUMI computing environment began in summer 2021, at which time an international team complemented with CSC experts installed the LUMI-C partition consisting of CPU processors. Acceptance and commissioning testing started immediately and were completed successfully. Watch the video: Construction of the LUMI data center and system installation.

 

In late 2021, LUMI-C was opened for pilot users. The purpose of piloting was to test the CPU partition and storage systems of the computer with heavy workloads. The 30 pilot projects selected from LUMI consortium countries represented several different disciplines. LUMI User Support Team conducted a comprehensive survey addressed to the pilot users, in which the user experience was found to be positive. Certain processes related to LUMI were also developed during the year.

Arrival of LUMI systemOne fifth of LUMI's capacity will be reserved for companies, and preparations for business use continued in 2021. Defining service paths was an important part of ensuring that the different use and funding models will be effective. The main focus in LUMI's business use will be on companies’ cooperation with higher education institutions and research institutes. This is the best way of bringing thematic research expertise and computing capacity for big data to bear on companies' RDI activities. LUMI may also be used directly by companies that have sufficient capabilities for this. In this case, the market price will be charged for using LUMI. CSC also worked with Business Finland. Using the Puhti AI platform and through Business Finland’s AI program, a few AI-driven computing projects were carried out, including Speechly’s project on developing speech recognition technology. The results of the projects were quite promising. Watch the video: Arrival of the LUMI system.

LUMI data centerIn late 2021, two major international awards added to the fame and reputation of LUMI. The first one, the annual Readers' and Editors' Choice Awards of the HPCwire magazine, was presented to LUMI for the Best Sustainability Innovation in HPC (HPCwire Editors’ Choice Award). The second international recognition was presented to LUMI at the global DCD Awards gala in London in December 2021, where LUMI received the Data Center Design Innovation Award together with Granlund and Synopsis Architects. Watch the video: LUMI data center.

The installation of the LUMI-G partition, which consists of LUMI GPU processors, will begin in spring 2022. This will be followed by a second piloting phase of the system, which aims to create heavy workloads for the GPUs and storage systems and test the scalability of the GPU partition. In addition, the purpose of the pilot phase is to receive feedback from pilot users. For this phase, pilot projects have been selected from each LUMI consortium country. The full LUMI installation will be available for use in autumn 2022. The Extreme Scale Access call for proposals was opened for Finnish projects aiming to use the LUMI-G partition in September 2021, and these projects will be launched as soon as LUMI-G is available to researchers.

New Funet network ready to respond to future needs

Funet, a fast and reliable information network for Finnish higher education institutions, research and education, covers all Finnish higher education institutions regardless of the administrative sector and serves over 390,000 end users throughout Finland.

Funet 2020, the Funet network's lifecycle update, was finished in early 2021. The update project was completed as planned. The new networking solution provides users with more reliable connectivity. Thanks to the update, the network will be highly scalable and capable of responding to future data transfer needs of education and research.

As proof of the successful completion of the Funet 2020 update project and the first-class data transfer capability of the upgraded network, 800 Gbit/s of data were transferred to a distance of more than 2,000 km on Funet in autumn 2021. This world record shows that Funet is highly scalable, ensuring that Finnish higher education institutions and research will have access to efficient data transfer capabilities well into the future. In addition, the network has excellent capabilities for responding to the future data transfer needs of EuroHPC LUMI and other key research resources.

CSC's research and project cooperation expanded further

CSC largely depends on competitive funding for developing its range of services. 

CSC participates in the Centre of Excellence in High Speed Electromechanical Energy Conversion Systems selected by the Academy of Finland for 2022–2029. Key reasons for CSC’s participation in the work of this Centre of Excellence is the ELMER software developed at CSC and special expertize in modelling multiphysics phenomena. 

NVIDIA, the Finnish Centre for Artificial Intelligence (FCAI) and CSC work together to run the NVAITC technology program, which promotes the capabilities of Finnish AI research by providing expertize and computing resources. The program cooperates with 40 research projects and has provided 11 projects with in-depth support in using graphics processors (GPU) more effectively to solve their research problems.

Together with the universities, CSC participates in two HPC projects approved by the Academy of Finland. The ExaFF project develops molecular dynamics software using efficient machine learning for LUMI and the other EuroHPC computers in a consortium led by Aalto University. The HISSA project of the University of Helsinki, the Sodankylä Geophysical Observatory and CSC aims to use LUMI and DL2021 infrastructure for efficient modelling and observation of ionosphere phenomena.

CSC is involved in the EU-funded E-TASC Helsinki Advanced Computing Hub, the particular aim of which is to improve the computational prerequisites of fusion energy. This project carried out together with VTT Technical Research Centre of Finland, Aalto University and Åbo Akademi University is the largest in Finland in the field of fusion research.

The Elements of Supercomputing open online course  intended to popularize high-performance computing was designed and implemented in cooperation with Kajaani University of Applied Sciences. The course was published at the DL2021 final seminar on 13 December 2021. The Finnish version of this open course was completed in early 2022. By mid-January 2022, 128 people had enrolled in the course and 25 had completed it. New MOOC content was produced to promote the more efficient use of CSC's computing resources and to support new users. This MOOC is also openly available to Finnish and international partners.

In 2021, we continued to build our activities in the area of sports and wellbeing together with our customers and partners. As a new project under the Kajaani Region Ecosystem Agreement, we launched a program for analysing sports and wellbeing data from Kainuu. The purpose of this program is to enable the collection of sports and wellbeing data into a data pool and its exploitation for various use cases, for example research purposes.

Back to top Go to Corporate Responsibility Report