Extracting and updating data performance during bulk numerical calculations in databases

Nowadays, more complex engineering problems need to carry out increasingly complex numerical calculations. In order to obtain the results as soon as possible, engineers must simplify the physical model on the one hand, apply appropriately efficient infrastructure to carry out calculations on the other. This infrastructure includes both hardware and software. Problems in numerical calculations in scientific applications are often caused by non-optimal front-end application code implementation or by ineffective system of scientific data management at back-end. In this article author presents some aspects of performance problems when relational database is used as backend storage. Of course, the biggest problem in numerical computation processing is that the large amount of data stored on the storage area can slow down the entire computing system and in turn directly affects the computational efficiency. Presented examples come from simulated OLTP (Online Transactional Processing) environments with large load and many queries executed. These examples can reflect real problems with data processing in numerical calculations on data extracted from MSSQL Server database with large storage system connected