Either SCHED_FIFO or SCHED_RR could be used; they are equivalent when the tasks consume less than a full slice.
We chose SCHED_FIFO because it was simpler; SCHED_RR has additional semantics and a hard-coded (from user space) timeslice that we didn't need. On Friday, August 9, 2013 11:26:20 PM UTC-7, Xin Qian wrote: > > Although the default time slice of SCHED_RR is 100ms, then FastMixer > could give up CPU before 100ms and allow other threads to be scheduled, so > what's the problem? Actually we've already done the experiment using > SCHED_RR, and the systrace result shows no difference. We couldn't figure > out the benefit of using FIFO policy. Thanks. > -- -- unsubscribe: android-porting+unsubscr...@googlegroups.com website: http://groups.google.com/group/android-porting --- You received this message because you are subscribed to the Google Groups "android-porting" group. To unsubscribe from this group and stop receiving emails from it, send an email to android-porting+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/groups/opt_out.
