High-performance computing processes trillions of data, models situations and changes climate predictions

Insight
2022 / 01 / 04
4 MIN. READING

Gytis Umantas, CEO of Novian Technologies

Today, it is hard to remember how we lived without a taxi, take-out food, online shopping, navigation, and many other applications. These services have created a more comfortable life, and their emergence has been driven by the availability and adaptability of advanced technologies.

This is the visible influence of modern technologies on everyday life. In addition, there are other activities where the applicability of technology and high-performance computing is also critical.

One such areas of science and research is weather forecasting and climate change research, as well as impact modelling. This area has been changing rapidly in recent times, and the main reason for this is technology: artificial intelligence, machine learning and high-performance computing.

The first high-performance computing cluster in Lithuania was installed in 2009 at the Vilnius University (VU) Computing Centre. For more than a decade, its services have been used by scientists from the faculties of Mathematics and Informatics, Physics and other faculties.

Today, the use of high performance computing (HPC) technologies is expanding further, with great potential in the life sciences, cyber simulations, weather forecasting, renewable energy and other areas.

New opportunities to apply these innovations in the country are provided by the Lithuania’s Economic Recovery and Resilience Plan, where one of the planned priorities is digitization and innovation, thus opening new opportunities for organisations to modernise their IT systems, which in turn strengthen the country’s economic growth potential.

Accurate and fast: supercomputer climate predictions

Why use high performance computing? First of all, with the help of this technology, the development of various models becomes highly accurate. This level of accuracy is relevant for assessing large arrays of historical and real-time data from different sources, such as weather forecasting, modelling the effects of climate change, or performing other data-intensive calculations.

Secondly, speed is also required to make such calculations. Accurate forecasts are possible by evaluating relevant data in real time, while a reliable and resilient data processing infrastructure is required to process data delivered every few minutes around the clock.

Why is accuracy and speed important in forecasting the weather and modelling the effects of climate change? Because a timely response to climate change can save lives, and its impact is also measurable financially.

Modelling the effects of climate change

In 2019, in Ouagadougou, the capital of Burkina Faso, Novian Technologies, the company offering IT infrastructure solutions, has designed and deployed a high-performance computing cluster. It addresses the challenges of modelling the effects of climate change, helping to prepare for change and mitigate its consequences.

During this project, we prepared the cluster architecture from scratch, supplied and installed the necessary equipment, configured the software, documented the system, and trained the client’s team. The installed HPC system consists of high-performance servers connected to a cluster of parallel computing using a high-performance and low-latency InfiniBand network.

These networks allow the collection of specific data on biophysical parameters, socio-economic indicators, meteorological, hydrological, land use and biological information. The different readings are automatically transmitted to the HPC system and processed. This simplifies the management of data collection, allows a faster and more efficient modelling of tasks, and prompt reporting.

In 2020, the year of the pandemic, we deployed under harsh conditions high-performance computing solutions in more than 10 climate-sensitive African capitals. They allow to analyse trillions of data per second, so data is collected, processed and shared much faster, forecasts are more accurate, and modelling results are obtained earlier. Generally speaking, the capacity is hundreds of thousands of times the capacity of a traditional computer.

As a result, agricultural development plans are more in line with projected climate change and ultimately help to avoid painful, including financial, consequences for local people. It is also important that local researchers, science and research centres can now work independently, relying less on the resources of other countries.

High-performance computing in Lithuania

The topic of high-performance computing is becoming increasingly relevant in Lithuania as well. In 2021, the implementation of a fundamentally modernised cluster of high-performance computing at the Faculty of Mathematics and Informatics of Vilnius University opened the possibilities to use the capacities of supercomputers not only for scientific but also for wider needs.

This architectural IT solution enables fast and efficient research in the fields of artificial intelligence development, neural network training, large-scale computing, and other creative and research activities that require an extremely large sample computing.

We have designed and implemented an innovative cluster of high-performance computing devices that meet the highest international standards in cooperation with customers and technology partners (DELL, IBM, NVIDIA and Mellanox).

We have also prepared a business plan, service catalogue and system accounting module for the university, which allows to offer a wide range of different technologies to the market, easily adapt it to the clients’ needs, promptly prepare the required environment and then perform the accounting. Thus, the HPC infrastructure of the open access centre can be used not only by researchers at VU, but also by researchers from other Lithuanian research institutions and companies.

Lithuania’s economic recovery plan also opens up new opportunities for IT modernisation

The Lithuania’s Economic Recovery and Resilience Plan approved in 2021 also provides opportunities to take advantage of the benefits offered by supercomputers.

Its priorities include digitisation and innovation, which opens up new opportunities for organisations to modernise their IT systems.

We also see the great potential of HPC technologies in the areas of life sciences, cyber simulations, weather forecasting, use of renewable energy sources, and other areas, and we are ready to help create value using HPC technologies with our knowledge and competencies.

Source: vz.lt. project