Hi all, I am currently working on my master thesis on a Linux project with regard to communication- and execution times. Therefore I want to trace some user-space applications using lttng-ust. Here I want to compute the best and worst case execution times and the response time of several threads. The simplest approach would be to trace the start and end of each execution, but this would neglect scheduling effects such as interference with higher priority tasks, where the thread would pause execution. What would be the best practice approach to calculate the CPU usage of each thread activation. Lttng seems to be doing something similar in their kernel traces.
I am using lttng-ust for tracing and pythonic babeltrace2 for parsing of the trace. Additionally it would be possible to collect an kernel trace if required. I really appreciate any help you can provide. Best regards Phil Hertha
_______________________________________________ lttng-dev mailing list lttng-dev@lists.lttng.org https://lists.lttng.org/cgi-bin/mailman/listinfo/lttng-dev
