Parallelized execution for maximum performance

Modern computer architectures allow parallel execution of various computations on the available CPU cores. It is also possible to transfer the computations to the graphics cards with the highest efficiency. Cyfex uses the latest technologies to combine all these features of computers for the maximum performance of an application.

Parallelization does come with a few problems, however: If the different computations are to access the same data, then this data access must be protected by mutual exclusion. We have developed a software module that solves this problem with optimal performance and with a guarantee. This module is integrated into real-time applications and supports thousands of end-clients in their work every day.

It is also often desirable that the individual elements are able to communicate with one another. Our class libraries enable this communication through object-oriented messages that are ideally integrated into an object-oriented software design.