Hello, I have a question regarding timing behaviour especially in SCHED_RR (use an OMAP3).
If there are running 2...5 jobs with same priority - how can I setup (and measure) a timeslice?
I checked some Kernel settings:
CONFIG_OMAP_32K_TIMER_HZ=128 CONFIG_NO_HZ=y CONFIG_HZ=128 CONFIG_TIMERFD=y # CONFIG_OMAP_MPU_TIMER is not set CONFIG_OMAP_32K_TIMER=y CONFIG_OMAP_32K_TIMER_HZ=128 CONFIG_OMAP_DM_TIMER=y CONFIG_HIGH_RES_TIMERS=y
Somewhere I read that the timeslice is 1/HZ -> 1/128 = 7.8ms ? Is this true, or what meaning does this CONFIG_HZ has?
Otherwise CONFIG_NO_HZ is set. What does this mean? How can I get a more predictable behaviour of the system?