Scientists of Novosibirsk State Technical University (NETI) continue to work on a digital workplace for users of supercomputing centers for collective use within the framework of the Priority 2030 program. Currently, three projects have been launched that will improve the infrastructure and expand the capabilities of researchers.
The digital workplace of users of supercomputer data centers is designed to significantly simplify the work of researchers with high-performance computing, systematize the accumulation of knowledge and open up new opportunities for interdisciplinary research not only for NSTU-NETI, but also for the region and the country as a whole.
According to Oleg Avrunev, director of the University's Informatization Center and head of the Digital Development Department at NSTU-NETI, the project, which was launched a year ago, is currently developing in two directions: supporting researchers at the stage of gaining access to computing resources and automating the efficient execution of complex calculations.
"One of the primary tasks of the project was to create a "single service window" for users who previously had to interact independently with various holders of computing resources. To date, a web application has been developed and is functioning for the simplified generation of applications for access to high-performance computer systems for students and staff of NSTU-NETI," Oleg Avrunev said.
The second, equally important component is the specialized EXCEL framework for computing automation. The tool allows researchers to set calculation tasks uniformly both on a personal computer and on supercomputers without routine operations. The system automatically handles file transfer, software environment configuration, and implementation of complex scenarios.
The user describes the task in the form of a graph showing the dependencies between operations. Such descriptions are systematically accumulated in a personal knowledge base and can be reused to solve new problems, which contributes to the accumulation and structuring of unique research techniques.
"The current version provides the ability to perform independent tasks in parallel within computational scenarios, which significantly speeds up the processing of large amounts of data. In addition, the observability and stability of the framework's executive subsystem have been improved, which makes its use more reliable," Oleg Avrunev emphasized.
The effectiveness of the platform has already been confirmed: the microseismic monitoring research group has developed its own library of data processing programs using the framework, compiled a number of computational scenarios, and successfully demonstrated parallel processing of data from daily monitoring of seismic activity in an oil field.
In parallel, research is underway to increase the framework's resilience to unstable Internet operation. The project team is studying computing management systems in various application areas to make decisions about the priorities of the framework's development. Also, a team of researchers from the Faculty of Applied Mathematics and Computer Science of NSTU-NETI uses the service to perform research work on optimizing customer logistics.
Work is actively underway to create an infrastructure that will empower researchers. Projects such as the development of an aggregator portal that will combine projects, solutions, developments and research in the field of artificial intelligence and high-performance computing at NSTU-NETI, a stand with large language models (LLM) and an aggregator portal for access to scientific equipment that allows students and staff to centrally provide information and manage access to scientific and production equipment of the university.