dims_create: fix calculation of factors for odd squares

[michaellass]

Until now sqrt(n) was missed as a factor for odd square numbers n. This lead to suboptimal results of MPI_Dims_create for input numbers like 9, 25, 49, ... Fix the results by including sqrt(n) in the search for factors.

Refs: #7186

[ompiteam-bot]

Can one of the admins verify this patch?

[awlauria]

ok to test

[michaellass]

Tested by creating 2D dims for numbers 1-99:

         Before         After  
1:     1       1      1      1
2:     2       1      2      1
3:     3       1      3      1
4:     2       2      2      2
5:     5       1      5      1
6:     3       2      3      2
7:     7       1      7      1
8:     4       2      4      2
9:     9       1      3      3   <---
10:    5       2      5      2
11:    11      1      11     1
12:    4       3      4      3
13:    13      1      13     1
14:    7       2      7      2
15:    5       3      5      3
16:    4       4      4      4
17:    17      1      17     1
18:    6       3      6      3
19:    19      1      19     1
20:    5       4      5      4
21:    7       3      7      3
22:    11      2      11     2
23:    23      1      23     1
24:    6       4      6      4
25:    25      1      5      5   <---
26:    13      2      13     2
27:    9       3      9      3
28:    7       4      7      4
29:    29      1      29     1
30:    6       5      6      5
31:    31      1      31     1
32:    8       4      8      4
33:    11      3      11     3
34:    17      2      17     2
35:    7       5      7      5
36:    6       6      6      6
37:    37      1      37     1
38:    19      2      19     2
39:    13      3      13     3
40:    8       5      8      5
41:    41      1      41     1
42:    7       6      7      6
43:    43      1      43     1
44:    11      4      11     4
45:    9       5      9      5
46:    23      2      23     2
47:    47      1      47     1
48:    8       6      8      6
49:    49      1      7      7   <---
50:    10      5      10     5
51:    17      3      17     3
52:    13      4      13     4
53:    53      1      53     1
54:    9       6      9      6
55:    11      5      11     5
56:    8       7      8      7
57:    19      3      19     3
58:    29      2      29     2
59:    59      1      59     1
60:    10      6      10     6
61:    61      1      61     1
62:    31      2      31     2
63:    9       7      9      7
64:    8       8      8      8
65:    13      5      13     5
66:    11      6      11     6
67:    67      1      67     1
68:    17      4      17     4
69:    23      3      23     3
70:    10      7      10     7
71:    71      1      71     1
72:    12      6      12     6
73:    73      1      73     1
74:    37      2      37     2
75:    15      5      15     5
76:    19      4      19     4
77:    11      7      11     7
78:    13      6      13     6
79:    79      1      79     1
80:    10      8      10     8
81:    9       9      9      9
82:    41      2      41     2
83:    83      1      83     1
84:    12      7      12     7
85:    17      5      17     5
86:    43      2      43     2
87:    29      3      29     3
88:    11      8      11     8
89:    89      1      89     1
90:    10      9      10     9
91:    13      7      13     7
92:    23      4      23     4
93:    31      3      31     3
94:    47      2      47     2
95:    19      5      19     5
96:    12      8      12     8
97:    97      1      97     1
98:    14      7      14     7
99:    11      9      11     9

[awlauria]

LGTM.

[jsquyres]

@michaellass Awesome! Would you mind opening a few more PRs -- git cherry-pick -x ...thiscommit... to the v3.0.x, v3.1.x, and v4.0.x branches? That will get this fix into the next releases of those series.

Thanks!