Performance considerations
In this feature article we will learn how to monitor performance for your applications and how to get around the most common bottlenecks.
Performance for numerical applications always comes along with the consumption of memory and - of course - with the efficiency of implementation. Both points get increased weight by increasing problem size. The bigger the problem, the stronger the need for
- fast executing algorithms, efficiently handling the data involved with the least instructions possible, and
- efficient memory management, limiting the computers memory used to the minimal amount possible in order to prevent for memory swapping by the system.
Read about the following topics:
