Hello, Paul. [Dropping CC]
> Hello, Joel, > > In case you are -seriously- interested... ;-) > > Thanx, Paul > > rcu_nocbs= > > Adding a CPU to this list offloads RCU callback invocation from > that CPU's softirq handler to a kthread. In big.LITTLE systems, > this kthread can be placed on a LITTLE CPU, which has been > demonstrated to save significant energy in benchmarks. > > http://www.rdrop.com/users/paulmck/realtime/paper/AMPenergy.2013.04.19a.pdf > I have checked our config. We do use rcu_nocbs=0-7 as kthreads but what i see those threads are not bound to 0-3 CPUs. In our case it is little cluster. I think i should check and run some test cases regarding power savings if i pin all threads to little cluster. > rcutree.rcu_idle_gp_delay= (Only CONFIG_RCU_FAST_NO_HZ=y kernels.) > > This defaults to four jiffies on the theory that grace periods > tend to last about that long. If grace periods tend to take > longer, then it makes a lot of sense to increase this. And maybe > battery-powered devices would rather have it be about 2x or 3x > the expected grace-period duration, who knows? > > I would keep it to a power of two, but the code should work with > other numbers. Except that I don't know that this has ever been > tested. ;-) > Same here. We do use it. -- Vlad Rezki
