rss-ladder: opportunity to ignore socket / numa node.

[strizhechenko]

lscpu -e

CPU NODE SOCKET CORE L1d:L1i:L2 ONLINE
0   0    0      0    0:0:0      да
1   0    1      1    1:1:1      да
2   0    0      2    2:2:0      да
3   0    1      3    3:3:1      да
4   0    0      4    4:4:2      да
5   0    1      5    5:5:3      да
6   0    0      6    6:6:2      да
7   0    1      7    7:7:3      да

lscpu -p

# The following is the parsable format, which can be fed to other
# programs. Each different item in every column has an unique ID
# starting from zero.
# CPU,Core,Socket,Node,,L1d,L1i,L2
0,0,0,0,,0,0,0
1,1,1,0,,1,1,1
2,2,0,0,,2,2,0
3,3,1,0,,3,3,1
4,4,0,0,,4,4,2
5,5,1,0,,5,5,3
6,6,0,0,,6,6,2
7,7,1,0,,7,7,3

BTW: it's more about binding to local biggest cache line. In this particular case there are 4 equal lines with 2 cpus and it's worse than just using all the cpus of the system. It's an old CPU:

# lscpu 
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                8
On-line CPU(s) list:   0-7
Thread(s) per core:    1
Core(s) per socket:    4
Socket(s):             2
NUMA node(s):          1
Vendor ID:             GenuineIntel
CPU family:            6
Model:                 23
Model name:            Intel(R) Xeon(R) CPU           E5410  @ 2.33GHz
Stepping:              6
CPU MHz:               2333.614
BogoMIPS:              4666.00
Virtualization:        VT-x
L1d cache:             32K
L1i cache:             32K
L2 cache:              6144K
NUMA node0 CPU(s):     0-7