High-performance computing offers companies a boost and competitive advantage

Machine learning is an area of artificial intelligence that efficiently automates the construction of analytical models. It can, for example, be used to support decision-making or to improve the efficiency of production processes. Machine learning algorithms identify patterns in data and help data scientists solve problems. Machine learning requires a large computing capacity.

In machine learning, algorithms are used to identify data models, which can then be used to generate predictions and forecasts. Machine learning algorithms can predict values, identify unusual events, define structures, and create categories. For example, machine learning can use customer data and identify behavioral models for optimizing product recommendations and providing the best possible customer experience.

Computational fluid dynamics deals with simulations of gas and liquid flows. It utilizes high-performance computing to produce a precise numerical solution for equations describing the movement of a liquid or gas. Computational fluid dynamics is needed in conducting basic research on fluid dynamics, designing complex flow configurations, and predicting interactions between chemical substances. Its application areas include aerodynamics, space and aviation research, weather simulations as well as engine and combustion analysis.

Our supercomputers have plenty of memory, and their high processing capacity makes it possible to turn data into meaningful information, output forecasts and recommendations. CSC has expertise in distributed data processing, data integration and data analytics. We can provide impartial advice on the best commercial and open source solutions available.

Natural language processing is an area of linguistics, computer science and artificial intelligence that deals with interactions between computers and human language, especially programming. The goal is to create a computer that can ‘understand’ document contents, including the contextual nuances of their language. This technology can accurately extract information and insights contained in documents as well as classify and organize documents.

A supercomputer is needed in natural language processing to analyze large volumes of natural language data. Its challenges often involve speech recognition as well as understanding and producing natural language.

Genomics is transitioning from a research-driven field to activities led by health care organizations. It has been predicted that health care providers will be producing the majority of genomic data in the near future. In order to maximize the value of genomic data, it is essential that data can be shared between organizations and across national borders. Research is strongly dependent on computing power, and our understanding of the genome has been identified as a major factor in health care development.

Drug development is a long process in which comprehensive computational methods can be used to focus expensive experimental phases on molecules that are the most likely to be effective. Atomic models can be used to study how medicinal substances bind with and affect target molecules or to test different hypotheses. While modeling has been used in the different research phases of pharmaceutical research for a long time, the capacity offered by supercomputers makes it possible to analyze enormous quantities of molecules at an unprecedented rate. In addition, the most sophisticated methods now produce accurate results from a smaller selected group of molecules at a faster rate than laboratory tests – and without the chemical waste.

The details of many chemical processes of industrial importance can be modeled using atomic-level methods, which include molecular dynamics and density functional theory. For example, simulations can be used to determine the operating mechanisms of catalysts as well as the properties of mixtures and solutions or their behavior on surfaces. This understanding can be used to create new, more efficient catalysts or to optimize process conditions. The power of supercomputers makes it possible to build and use models that have significance in the real world.

Quantum computers are revolutionizing research and development that utilizes computational modelling. Due to a completely new way of computing, quantum computers can solve some problems much more efficiently than classical supercomputers. In order to take advantage of the computational power of quantum computers, problems must first be reformulated in a form that is amenable for quantum computing. Particularly suitable applications for quantum computers are, e.g., financial modelling, risk analysis, machine learning, molecular and materials sciences, as well as logistics and optimization problems in general. Together with VTT and Aalto University, CSC maintains and develops the Finnish Quantum-Computing Infrastructure FiQCI. FiQCI harnesses the combined power of supercomputers and quantum computers, and provides our customers with a development and testing platform for quantum computing.

When the project results are the property of the company rather than public, the company concludes a contract on using the computing services with CSC and pays the fee set out on the service price list for using LUMI’s resources.

When the RD project is led by a research institute or higher education institution:

• LUMI’s resources can be used free of charge if the project results are published in keeping with scientific practice.
• If the results are not public, the fee set out on the service price list will be charged for using LUMI.

In an RDI project funded by Business Finland:

• When the value of the project is between EUR 20,000 and 80,000, a Finnish start-up or SME may access free computing capacity.
• Large and medium-sized enterprises can include LUMI’s computing services in their project budget and receive funding for at most 40% of the costs.

The value of the computing capacity is determined according to the service price list.

The European High Performance Computing Joint Undertaking’s (EuroHPC JU) has opened a continuous call for AI and Data-Intensive Applications. The call is intended to serve industry organizations, small to medium enterprises (SMEs), startups, as well as public sector entities, requiring access to supercomputing resources to perform artificial intelligence and data-intensive activities.

The call aims to support ethical artificial intelligence, machine learning, and in general, data-intensive applications, with a particular focus on foundation models and generative AI, e.g. large language models.

The EuroHPC supercomputers provided for this call are LUMI, MareNostrum5, Leonardo, Meluxina, Karolina and Vega.

The call is continuously open, with pre-defined cut-off dates for proposals. Read more about the call from EuroHPC JU homepage.

The company can start using the LUMI computing service in a ‘Try&Buy’ project, in which you can try out LUMI’s suitability for your purpose and test it before making the investment. 

A test project includes

• a LUMI project and credentials for two users
• CPU, GPU, and storage resources for testing
• expert support for LUMI project deployment.

A test project will be available for a limited period of time. After the test project, the company can continue using LUMI by signing a contract for the use of computing services with CSC and purchasing computing resources for an existing project for a fee set out on the service price list. Alternatively, the company may decide to give up using the supercomputer after the test phase without incurring charges.