Question about scheduler load balancing

Hello Presumably the scheduler load balancing logic is less reluctant to load balance tasks at higher level scheduling domains (e.g. cross NUMA) than lower level SD Is that true? If yes, which knobs (or hardcodes) drive the load_balance() function to vary how aggressively it balances when it is in

Re: Question about group scheduler cpu shares

2.How does that work when an interior node has multiple tasks? So if I added tasks E and F with difference nice levels to Group 1 task list how would the math work? Appreciate the help -thanks! On Thu, Jul 21, 2016 at 7:51 PM, Mike Galbraith wrote: > On Thu, 2016-07-21 at 18:18 -0700, Gaurav Pooth

Re: Question about group scheduler cpu shares

+Ingo, Peter On Thu, Jul 21, 2016 at 12:03 PM, Gaurav Poothia wrote: > Hi, > The cgroup "cpu" subsystem's weight calculation using "cpu shares" is > fairly clear when all tasks are attached to the leaf cgroups in "cpu" > subsystem. > For examp

Question about group scheduler cpu shares

Hi, The cgroup "cpu" subsystem's weight calculation using "cpu shares" is fairly clear when all tasks are attached to the leaf cgroups in "cpu" subsystem. For example in this cgroup hierarchy with tasks A.B,C and D and associated cpu shares: ROOT | + -Group1(3072) || | +- A(2048) || |