|
1 # TestVector definition file for perfmetrics.exe |
|
2 |
|
3 # This testvector is specifically suited for doing ONLY search types |
|
4 # of tests on BIG indexes like poi. To be able to run the tests you |
|
5 # actually need a one-liner test corpus from which one record is read |
|
6 # but is NOT added actually to the index (number of firsts is -1). |
|
7 |
|
8 # name of the test. If using 'poi', the path |
|
9 # data\\indexing\\searched\\perfmetrics\\pois.txt is used instead of |
|
10 # HugeTestCorpus as well as POI schema applied to base app class "root |
|
11 # maps poi". |
|
12 #plain |
|
13 poi |
|
14 |
|
15 # (re-)create index database on start (1: yes, 0: no) |
|
16 #1 |
|
17 0 |
|
18 |
|
19 # RAM Buffer Size (in bytes) |
|
20 32768 |
|
21 |
|
22 # number of preindexed (or skipped) items |
|
23 # if positive, this many items are read from corpus, indexing |
|
24 # if negative, this many items are read from corpus, discarding |
|
25 0 |
|
26 |
|
27 # max input size - this many items are read from the corpus at most |
|
28 # (not counting the preindexed items) |
|
29 #60000 |
|
30 #147729 |
|
31 1 |
|
32 |
|
33 # lump size - measurements are organized into lumps |
|
34 #5000 |
|
35 1 |
|
36 |
|
37 # number of firsts - for each lump, this many items are indexed (while |
|
38 # measuring time and memory consumption) |
|
39 #10 |
|
40 -1 |
|
41 |
|
42 |
|
43 # number of lasts - for each lump, after indexing the firsts |
|
44 # (measuring them), and possibly (usually) indexing more, the last |
|
45 # items are also of special interest (they are added, deleted, |
|
46 # re-added, updated: all this while taking measurements) |
|
47 #10 |
|
48 0 |
|
49 |
|
50 # searches - for each query term, this many times the search is |
|
51 # executed in a loop (can be 0 to switch of search measurements) |
|
52 #5 |
|
53 5 |
|
54 |
|
55 # incremental searches - for each query term, this many times an |
|
56 # incremental search is executed in a loop (can be 0 to switch of |
|
57 # incremental search measurements) |
|
58 #5 |
|
59 5 |
|
60 |
|
61 # minimal incremental length: the minimal perfix length for each query |
|
62 # term to form a wildcard query for incremental search |
|
63 # purposes. E.g. if the query term is "character" and the minimal |
|
64 # incremental length is 2, then wildcard searches start with "ch*" |
|
65 3 |
|
66 |
|
67 # suggestion searches - for each query term, this many times an |
|
68 # suggest searches is executed in a loop (can be 0 to switch of |
|
69 # suggest search measurements) |
|
70 5 |
|
71 |
|
72 # minimal suggestion length: the minimal perfix length for each query |
|
73 # term to form a wildcard query for suggestion |
|
74 # purposes. E.g. if the query term is "character" and the minimal |
|
75 # incremental length is 2, then wildcard searches start with "ch*" |
|
76 2 |
|
77 |
|
78 # whether to take snapshot of the index at certain points (copying the |
|
79 # physical index to c:\data\perfmetrics\idxs\NNN\... where NNN is the |
|
80 # id of the lump at the end of which the snapshot was taken. Values: 0 |
|
81 # no, 1 yes. |
|
82 #0 |
|
83 0 |
|
84 |
|
85 # whether to add lasts. If "number of lasts" > 0 and "add lasts" == 1 |
|
86 # (true) then the lasts are added. 0: no, 1: yes |
|
87 1 |
|
88 |
|
89 # whether to delete and re-add lasts. If "number of lasts" > 0 and |
|
90 # "delete and re-add lasts" == 1 (true) then the lasts are deleted and |
|
91 # re-added. 0: no, 1: yes |
|
92 1 |
|
93 |
|
94 # whether to update lasts. If "number of lasts" > 0 and "update lasts" |
|
95 # == 1 (true) then the lasts are updated. 0: no, 1: yes |
|
96 1 |
|
97 |
|
98 # query terms - these terms will be searched for at the end of every |
|
99 # lump |
|
100 bmp |
|
101 xiao |
|
102 cheng |
|
103 John |
|
104 Neilson |
|
105 Hilda |