1 ============================
2 NUMA resource associativity
3 ============================
4
5 Associativity represents the groupings of the
6 domains of substantially similar mean performa
7 of that domain. Resources subsets of a given d
8 performance relative to each other than relati
9 are represented as being members of a sub-grou
10 characteristic is presented in terms of NUMA n
11 From the platform view, these groups are also
12
13 PAPR interface currently supports different wa
14 grouping details to the OS. These are referred
15 associativity grouping. Form 0 is the oldest f
16
17 Hypervisor indicates the type/form of associat
18 Bit 0 of byte 5 in the "ibm,architecture-vec-5
19 A value of 1 indicates the usage of Form 1 ass
20 bit 2 of byte 5 in the "ibm,architecture-vec-5
21
22 Form 0
23 ------
24 Form 0 associativity supports only two NUMA di
25
26 Form 1
27 ------
28 With Form 1 a combination of ibm,associativity
29 device tree properties are used to determine t
30
31 The “ibm,associativity” property contains
32 representing the resource’s platform groupin
33
34 The “ibm,associativity-reference-points” p
35 (domainID index) that represents the 1 based o
36 The list of domainID indexes represents an inc
37
38 ex:
39 { primary domainID index, secondary domainID i
40
41 Linux kernel uses the domainID at the primary
42 Linux kernel computes NUMA distance between tw
43 if they belong to the same higher-level domain
44 level of the resource group, the kernel double
45 comparing domains.
46
47 Form 2
48 -------
49 Form 2 associativity format adds separate devi
50 thereby making the node distance computation f
51 domain numbering. With numa distance computati
52 "ibm,associativity-reference-points" property,
53 ids at the same domainID index representing re
54 characteristics.
55
56 Hypervisor indicates the usage of FORM2 associ
57 "ibm,architecture-vec-5" property.
58
59 "ibm,numa-lookup-index-table" property contain
60 the domainIDs present in the system. The offse
61 used as an index while computing numa distance
62
63 prop-encoded-array: The number N of the domain
64 N domainID encoded as with encode-int
65
66 For ex:
67 "ibm,numa-lookup-index-table" = {4, 0, 8, 250
68 computing the distance of domain 8 from other
69 this document, this offset will be referred to
70
71 "ibm,numa-distance-table" property contains a
72 distance between resource groups/domains prese
73
74 prop-encoded-array: The number N of the distan
75 N distance values encoded as with encode-bytes
76 The number N must be equal to the square of m
77 numa-lookup-index-table.
78
79 For ex:
80 ibm,numa-lookup-index-table = <3 0 8 40>;
81 ibm,numa-distace-table = <9>, /bits/ 8 < 10 2
82
83 ::
84
85 | 0 8 40
86 --|------------
87 |
88 0 | 10 20 80
89 |
90 8 | 20 10 160
91 |
92 40| 80 160 10
93
94 A possible "ibm,associativity" property for re
95
96 { 3, 6, 7, 0 }
97 { 3, 6, 9, 8 }
98 { 3, 6, 7, 40}
99
100 With "ibm,associativity-reference-points" { 0
101
102 "ibm,lookup-index-table" helps in having a com
103 Since domainID can be sparse, the matrix of di
104 With "ibm,lookup-index-table" we can achieve a
105 distance information.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.