The goal of the thesis is to develop a tool for graphic interpretation of results of kernel performance measurement. The tool should also be able to automatically detect and report performance regression. The student has to study existing benchmarks for measuring the performance of Linux kernel scheduler and means of storing of the benchmark results.
- Get acquainted with the existing methods for measuring performance of the Linux kernel scheduler and with means of storing of benchmarks results for further processing.
- Study possible ways of processing these results with a focus on graphic interpretation and on methods for detection of performance degradation.
- Design and implement a method for efficient graphic interpretation of long-term measurements.
- Design and implement a method for automatic detection of performance regression.
- Demonstrate the functionality of your implementation on at least two versions of the Linux kernel.
- Evaluate the obtained results and discuss possibilities of further development of the project, especially of the automatic detection of performance regression.