Specific setting and detailed vm and core corresponding can see the Readme-v6.md and Readme-v8.md.
The firecrackers and containers in Readme-v6.md have the same setting, each vm set to one cpu core.
The firecreackers in Readme-v8.md each vm set to one, two or more cores.
Use mpstat -P ALL 1 to generate "cpu_perf.txt".
Use fio_extract.py to deal data and get graphs.
- Graphs:
- mpstat_all.png
- mpstat_UsrSysEtc.png
- mpstat_data_overtime.png (cpu utilization overtime)
mpstat_data_overtime.png is the same as the excels in folder v6's randomCore-v6, setOddEvenCore-v6 and folder v8's randomCore-v8, setOddEvenCore-v8.
Use cat /proc/schedstat to generate "schedstat.txt".
Use schedstat_extract.py to deal data and get graphs.
Get the colume 7, 8, 9 of the cpu0 to 31
- Meaning
- Colume 9 means: # of tasks (not necessarily unique) given to the processor or means: # of timeslices run on this cpu.
- Can refer to http://eaglet.pdxhosts.com/rick/linux/schedstat/v15/format-15.html or https://docs.kernel.org/scheduler/sched-stats.html for detailed info of colume 7, colume 8.
- Graphs:
- schedstat_fig_col_7.png
- schedstat_fig_col_8.png
- schedstat_fig_col_9.png
Use cat /proc/sched_debug to generate "schedDebug.txt".
Use schedDebug_extract.py to deal data and get graphs.
Get the .nr_switches of each core.
- Graphs:
- schedDebug_fig.png
Use cat /proc/interrupts to generate "interrupts.txt".
Use interrupts_extract.py to deal data and get graphs.
Get the RES: ....... Rescheduling interrupts of each core.
- Graphs:
- interrupts_fig.png