Skip to content

RFC sched: Fix scheduling for multi-node processors

Andreas Herrmann lkml 1 version In review

Get series with b4
b4 am [email protected]

Versions newest first

  1. v1 RFC Latest15 patches 48 repliesIn review

Patches in v1 No applies recorded yet

#Subject and filesSizeRepliesApplied
1/15x86, sched: Add config option for multi-node CPU schedulingarch/x86/Kconfig+10 -02 replies
2/15sched, x86: Provide initializer for MN scheduling domain, define MN levelarch/x86/include/asm/topology.h include/linux/sched.h+26 -00 replies
3/15sched: Add cpumask to be used when building MN domainkernel/sched.c+7 -10 replies
4/15sched: Define per CPU variables and cpu_to_group function for MN domainkernel/sched.c+27 -50 replies
5/15sched: Add function to build MN sched domainkernel/sched.c+21 -00 replies
6/15sched: Add support for MN domain in build_sched_groupskernel/sched.c+10 -00 replies
7/15sched: Activate build of MN domainskernel/sched.c+8 -00 replies
8/15sched: Add parameter sched_mn_power_savings to control MN domain sched policyinclude/linux/sched.h kernel/sched.c+35 -710 replies
9/15sched: Check sched_mn_power_savings when setting flags for CPU and MN domainsarch/x86/include/asm/topology.h include/linux/sched.h+14 -33 replies
10/15sched: Check for sched_mn_power_savings when doing load balancingkernel/sched.c+3 -23 replies
11/15sched: Pass unlimited __cpu_power information to upper domain level groupsinclude/linux/sched.h kernel/sched.c+31 -138 replies
12/15sched: Allow NODE domain to be parent of MC instead of CPU domainkernel/sched.c+73 -332 replies
13/15sched: Detect child domain of NUMA (aka NODE) domainkernel/sched.c+41 -22 replies
14/15sched: Conditionally limit __cpu_power when child sched domain has type NODEkernel/sched.c+2 -22 replies
15/15x86: Fix cpu_coregroup_mask to return correct cpumask on multi-node processorsarch/x86/kernel/smpboot.c+1 -116 replies

Reviews trailers per patch

No Reviewed-by, Acked-by or Tested-by trailers for v1 yet.

Read the v1 thread

Cover letter [RFC][PATCH 0/15] sched: Fix scheduling for multi-node processors

Hi,

Subsequent patches adapt scheduling code to support multi-node processors.

In short, the required changes need to fulfill two requirements:

(1) The set of CPUs in a NUMA node does not necessarily span CPUs of entire sockets anymore. (Current code assumes that.)

(2) The additional hierarchy in the CPU topology (i.e. internal node) might be useful when doing load balancing when power saving matters.

Patches 1-7 (add basic) support fo a new scheduling domain (called MN for multi-node) Patch 8 adds a knob to control power_savings balancing for MN domain Patches 9, 10 add the snippets to do the power_savings balancing for MN domain Patch 11 adds a way to pass unlimited __cpu_power information to upper domain levels Patch 12 allows NODE domain to be parent of MC domain (and thus child of MN domain) Patch 13 detects whether NODE domain is parent of MC instead of CPU domain Patch 14 fixes perf policy scheduling when NODE domain is parent of MC domain Patch 15 fixes cpu_coregroup_mask to use mask of core_siblings instead of node_siblings
(I admit that this change should better be added to my topology patches.)

To apply the patches you need to use tip/master as of today (containing the sched cleanup patches) plus the 8 topology patches that I have sent recently. (See
http://marc.info/?l=linux-kernel&m=124964980507887)

Full power saving scheduling options on multi-node processors are only available with CONFIG_SCHED_MN=y. See example and example output below.

Regards,

Andreas

PS: I send this as RFC. It seems to be pretty stable, though. But I
want to do some more testing before I'd ask to apply this to tip-tree and I also want to check the performance impact of the new sched domain level.

Signature · 272 lines
Operating | Advanced Micro Devices GmbH
  System  | Karl-Hammerschmidt-Str. 34, 85609 Dornach b. München, Germany
 Research | Geschäftsführer: Thomas M. McCoy, Giuliano Meroni
  Center  | Sitz: Dornach, Gemeinde Aschheim, Landkreis München
  (OSRC)  | Registergericht München, HRB Nr. 43632
--------------------------------------------------------------------------------
Examples:
=========
(1) To demonstrate power_savings balancing I provide top output when system
    is partially loaded.

 (Note: sched_mc_power_savings=sched_mn_power_savings=0)

 # for i in `seq 1 6`; do nbench& done

 top - 15:41:08 up 7 min,  3 users,  load average: 2.49, 0.64, 0.21
 Tasks: 267 total,   7 running, 260 sleeping,   0 stopped,   0 zombie
 Cpu0  : 49.7%us,  0.0%sy,  0.0%ni, 50.3%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu1  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu2  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu3  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu4  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu5  :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu6  : 50.2%us,  0.0%sy,  0.0%ni, 49.8%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu7  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu8  :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu9  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu10 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu11 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu12 : 48.8%us,  0.0%sy,  0.0%ni, 51.2%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu13 :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu14 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu15 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu16 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu17 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu18 : 50.8%us,  0.0%sy,  0.0%ni, 49.2%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu19 :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu20 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu21 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu22 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu23 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st

 # echo 1  >> /sys/devices/system/cpu/sched_mn_power_savings

 top - 15:42:27 up 8 min,  3 users,  load average: 3.91, 1.49, 0.54
 Tasks: 267 total,   7 running, 260 sleeping,   0 stopped,   0 zombie
 Cpu0  :  0.0%us,  0.3%sy,  0.0%ni, 99.7%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu1  :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu2  :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu3  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu4  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu5  :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu6  :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu7  :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu8  :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu9  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu10 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu11 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu12 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu13 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu14 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu15 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu16 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu17 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu18 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu19 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu20 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu21 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu22 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu23 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st

 # echo 2  >> /sys/devices/system/cpu/sched_mc_power_savings 

 top - 15:43:09 up 9 min,  3 users,  load average: 4.93, 2.06, 0.77
 Tasks: 267 total,   7 running, 260 sleeping,   0 stopped,   0 zombie
 Cpu0  : 99.0%us,  0.0%sy,  0.0%ni,  1.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu1  :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu2  :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu3  :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu4  :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu5  :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu6  :  0.7%us,  0.0%sy,  0.0%ni, 99.3%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu7  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu8  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu9  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu10 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu11 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu12 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu13 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu14 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu15 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu16 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu17 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu18 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu19 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu20 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu21 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu22 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu23 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st

 # echo 0  >> /sys/devices/system/cpu/sched_mc_power_savings 
 # echo 0  >> /sys/devices/system/cpu/sched_mn_power_savings 

 top - 15:44:22 up 10 min,  3 users,  load average: 5.38, 2.86, 1.15
 Tasks: 267 total,   7 running, 260 sleeping,   0 stopped,   0 zombie
 Cpu0  : 49.2%us,  0.3%sy,  0.0%ni, 50.5%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu1  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu2  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu3  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu4  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu5  :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu6  : 50.8%us,  0.0%sy,  0.0%ni, 49.2%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu7  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu8  :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu9  :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu10 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu11 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu12 : 49.2%us,  0.0%sy,  0.0%ni, 50.8%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu13 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu14 :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu15 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu16 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu17 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu18 : 50.7%us,  0.0%sy,  0.0%ni, 49.3%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu19 :100.0%us,  0.0%sy,  0.0%ni,  0.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu20 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu21 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu22 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st
 Cpu23 :  0.0%us,  0.0%sy,  0.0%ni,100.0%id,  0.0%wa,  0.0%hi,  0.0%si,  0.0%st

--------------------------------------------------------------------------------
(2) To illustrate the new domain hierarchy I give examples for sched
    domains and groups of CPU 23 of my test system:

CONFIG_SCHED_MN=n
sched_mc_power_savings=0

 CPU23 attaching sched-domain:
  domain 0: span 18-23 level MC
   groups: 23 18 19 20 21 22
   domain 1: span 18-23 level NODE
    groups: 18-23 (__cpu_power = 6144)
    domain 2: span 0-23 level CPU
     groups: 18-23 0-5 6-11 12-17

CONFIG_SCHED_MN=n
sched_mc_power_savings=2

 CPU23 attaching sched-domain:
  domain 0: span 18-23 level MC
   groups: 23 18 19 20 21 22
   domain 1: span 18-23 level NODE
    groups: 18-23 (__cpu_power = 6144)
    domain 2: span 0-23 level CPU
     groups: 18-23 (__cpu_power = 6144) 0-5 (__cpu_power = 6144) 6-11 (__cpu_power = 6144) 12-17 (__cpu_power = 6144)

CONFIG_SCHED_MN=y, CONFIG_NUMA=y
sched_mc_power_savings=0, sched_mn_power_savings=0

 CPU23 attaching sched-domain:
  domain 0: span 18-23 level MC
   groups: 23 18 19 20 21 22
   domain 1: span 18-23 level NODE
    groups: 18-23 (__cpu_power = 6144)
    domain 2: span 12-23 level MN
     groups: 18-23 12-17
     domain 3: span 0-23 level CPU
      groups: 12-23 0-11

CONFIG_SCHED_MN=y, CONFIG_NUMA=y
sched_mc_power_savings=0, sched_mn_power_savings=1

 CPU23 attaching sched-domain:
  domain 0: span 18-23 level MC
   groups: 23 18 19 20 21 22
   domain 1: span 18-23 level NODE
    groups: 18-23 (__cpu_power = 6144)
    domain 2: span 12-23 level MN
     groups: 18-23 12-17
     domain 3: span 0-23 level CPU
      groups: 12-23 (__cpu_power = 12288) 0-11 (__cpu_power = 12288)

CONFIG_SCHED_MN=y, CONFIG_NUMA=y
sched_mc_power_savings=2, sched_mn_power_savings=0

 CPU23 attaching sched-domain:
  domain 0: span 18-23 level MC
   groups: 23 18 19 20 21 22
   domain 1: span 18-23 level NODE
    groups: 18-23 (__cpu_power = 6144)
    domain 2: span 12-23 level MN
     groups: 18-23 (__cpu_power = 6144) 12-17 (__cpu_power = 6144)
     domain 3: span 0-23 level CPU
      groups: 12-23 (__cpu_power = 12288) 0-11 (__cpu_power = 12288)

CONFIG_SCHED_MN=y, CONFIG_NUMA=y, CONFIG_ACPI_NUMA=n
(and CONFIG_SCHED_MN=y, CONFIG_NUMA=n)
sched_mc_power_savings=0, sched_mn_power_savings=0

 CPU23 attaching sched-domain:
  domain 0: span 18-23 level MC
   groups: 23 18 19 20 21 22
   domain 1: span 12-23 level MN
    groups: 18-23 12-17
    domain 2: span 0-23 level CPU
     groups: 12-23 0-11

CONFIG_SCHED_MN=y, CONFIG_NUMA=y, CONFIG_ACPI_NUMA=n
(and CONFIG_SCHED_MN=y, CONFIG_NUMA=n)
sched_mc_power_savings=0, sched_mn_power_savings=1

 CPU23 attaching sched-domain:
  domain 0: span 18-23 level MC
   groups: 23 18 19 20 21 22
   domain 1: span 12-23 level MN
    groups: 18-23 12-17
    domain 2: span 0-23 level CPU
     groups: 12-23 (__cpu_power = 12288) 0-11 (__cpu_power = 12288)

CONFIG_SCHED_MN=y, CONFIG_NUMA=y, CONFIG_ACPI_NUMA=n
(and CONFIG_SCHED_MN=y, CONFIG_NUMA=n)
sched_mc_power_savings=2, sched_mn_power_savings=0

 CPU23 attaching sched-domain:
  domain 0: span 18-23 level MC
   groups: 23 18 19 20 21 22
   domain 1: span 12-23 level MN
    groups: 18-23 (__cpu_power = 6144) 12-17 (__cpu_power = 6144)
    domain 2: span 0-23 level CPU
     groups: 12-23 (__cpu_power = 12288) 0-11 (__cpu_power = 12288)

--------------------------------------------------------------------------------
(3) Further information -- just for completeness.

With NUMA support and SRAT detection the kernel uses following NUMA
information:

  # numactl --hardware
 available: 4 nodes (0-3)
 node 0 cpus: 0 1 2 3 4 5
 node 0 size: 2047 MB
 node 0 free: 1761 MB
 node 1 cpus: 6 7 8 9 10 11
 node 1 size: 2046 MB
 node 1 free: 1990 MB
 node 2 cpus: 18 19 20 21 22 23
 node 2 size: 2048 MB
 node 2 free: 2004 MB
 node 3 cpus: 12 13 14 15 16 17
 node 3 size: 2048 MB
 node 3 free: 2002 MB
 node distances:
 node   0   1   2   3 
   0:  10  16  16  16 
   1:  16  10  16  16 
   2:  16  16  10  16 
   3:  16  16  16  10 

Without ACPI SRAT support (e.g. CONFIG_ACPI_NUMA=n) the NUMA
information is:

 # numactl  --hardware
 available: 1 nodes (0-0)
 node 0 cpus: 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
 node 0 size: 8189 MB
 node 0 free: 7900 MB
 node distances:
 node   0 
   0:  10

--------------------------------------------------------------------------------
FINI

Read the cover letter in its thread