Wiki source code of Performance test on Jetty/HSQLDB with a single wiki
Version 62.2 by Thomas Mortagne on 2014/10/02 16:33
Show last authors
| author | version | line-number | content |
|---|---|---|---|
| 1 | {{toc/}} | ||
| 2 | |||
| 3 | = Environment = | ||
| 4 | |||
| 5 | * Ubuntu 14.04 64b | ||
| 6 | * Quad core i7 | ||
| 7 | * Oracle JDK 1.8.0_20 64b | ||
| 8 | * standard jetty+hsqldb | ||
| 9 | * only main wiki | ||
| 10 | |||
| 11 | = Jetty startup = | ||
| 12 | |||
| 13 | (% style="color:orange" %)(/) Same | ||
| 14 | |||
| 15 | Unit: seconds | ||
| 16 | |||
| 17 | Until end message is printer. Done by hand with a chronometer. | ||
| 18 | |||
| 19 | * 5.4.5: 12 | ||
| 20 | * 6.2: 12 | ||
| 21 | |||
| 22 | = First access = | ||
| 23 | |||
| 24 | First HTTP request after starting Jetty. | ||
| 25 | |||
| 26 | == First access to not existing page without UI (/xwiki/bin/get/NoSpace/NoPage). == | ||
| 27 | |||
| 28 | (% style="color:orange" %)(x) Same | ||
| 29 | |||
| 30 | Unit: seconds | ||
| 31 | |||
| 32 | Firefox network statistics to load the root resource (image/js/css/etc. resources are not taken into account). | ||
| 33 | |||
| 34 | * 5.4.5 : 6 | ||
| 35 | * 6.2 : 6 | ||
| 36 | |||
| 37 | == First access to not existing page with UI (/xwiki/bin/get/NoSpace/NoPage). == | ||
| 38 | |||
| 39 | (% style="color:red" %)(x) 3 times slower | ||
| 40 | |||
| 41 | Unit: seconds | ||
| 42 | |||
| 43 | Firefox network statistics to load the root resource (image/js/css/etc. resources are not taken into account). | ||
| 44 | |||
| 45 | * 5.4.5 : 6 | ||
| 46 | |||
| 47 | Done by hand with a chronometer (image/js/css/etc. resources are not taken into account). | ||
| 48 | |||
| 49 | * 6.2 : 18 | ||
| 50 | |||
| 51 | = Reload = | ||
| 52 | |||
| 53 | XWiki is doing nothing on background and everything is both in server and client cache already. Firefox network statistics to load the root resource (image/js/css/etc. resources are not taken into account). | ||
| 54 | |||
| 55 | == Reload of not existing page without UI (/xwiki/bin/get/NoSpace/NoPage) == | ||
| 56 | |||
| 57 | (% style="color:red" %)(x) more that 2 times slower | ||
| 58 | |||
| 59 | Unit: milliseconds | ||
| 60 | |||
| 61 | * 5.4.5 : [10, 20] | ||
| 62 | * 6.2 : [30, 40] | ||
| 63 | |||
| 64 | == Reload of not existing page with UI (/xwiki/bin/view/NoSpace/NoPage) == | ||
| 65 | |||
| 66 | (% style="color:red" %)(x) 2 times slower | ||
| 67 | |||
| 68 | Unit: milliseconds | ||
| 69 | |||
| 70 | * 5.4.5 : [100, 150] | ||
| 71 | * 6.2 : [200, 300] | ||
| 72 | |||
| 73 | == Reload of Main.WebHome with UI == | ||
| 74 | |||
| 75 | (% style="color:red" %)(x) 2 times slower | ||
| 76 | |||
| 77 | Unit: milliseconds | ||
| 78 | |||
| 79 | * 5.4.5 : [250, 350] | ||
| 80 | * 6.2 : [400, 500] | ||
| 81 | |||
| 82 | == Reload of Main.WebHome without UI == | ||
| 83 | |||
| 84 | (% style="color:red" %)(x) more than 2 times slower | ||
| 85 | |||
| 86 | Unit: milliseconds | ||
| 87 | |||
| 88 | * 5.4.5 : [65, 90] | ||
| 89 | * 6.2 : [145, 190] | ||
| 90 | |||
| 91 | = SOLR = | ||
| 92 | |||
| 93 | == Full SOLR reindex == | ||
| 94 | |||
| 95 | Unit: seconds | ||
| 96 | |||
| 97 | Delete the data/solr folder, restart, load home page and calculate the time of activity of the SOLR index thread using Yourkit (Yourkit is slowing down a bit all that but what is important is the difference anyway). | ||
| 98 | |||
| 99 | * 5.4.5 : 51 | ||
| 100 | * 6.2 : {{todo/}} | ||
| 101 | |||
| 102 | == SOLR sync == | ||
| 103 | |||
| 104 | Only the thread determining what need to be indexed and not the actual indexing. | ||
| 105 | |||
| 106 | === when index is empty === | ||
| 107 | |||
| 108 | (% style="color:orange" %)(/) Same | ||
| 109 | |||
| 110 | Unit: seconds | ||
| 111 | |||
| 112 | * 5.4.5 : 1 | ||
| 113 | * 6.2 : 1 | ||
| 114 | |||
| 115 | === when there is nothing to do === | ||
| 116 | |||
| 117 | (% style="color:orange" %)(/) Same | ||
| 118 | |||
| 119 | Unit: seconds | ||
| 120 | |||
| 121 | * 5.4.5 : < 1 | ||
| 122 | * 6.2 : < 1 | ||
| 123 | |||
| 124 | = Rendering = | ||
| 125 | |||
| 126 | == Page with 1000 macros without UI == | ||
| 127 | |||
| 128 | (% style="color:green" %)(/) almost 3 times quicker | ||
| 129 | |||
| 130 | 1000 times ~{~{id}} macro, one on each line. | ||
| 131 | |||
| 132 | Unit: milliseconds | ||
| 133 | |||
| 134 | * 5.4.5 : 9500 | ||
| 135 | * 6.2 : 3500 | ||
| 136 | |||
| 137 | = Memory = | ||
| 138 | |||
| 139 | Attach Yourkit to get the information. | ||
| 140 | |||
| 141 | Unit: MB | ||
| 142 | |||
| 143 | == Memory after jetty startup == | ||
| 144 | |||
| 145 | (% style="color:green" %)(/) use and keep less | ||
| 146 | |||
| 147 | Start jetty and wait until java is not doing anything on both (wait for potential initialization of various stuff like core extension cache, etc). | ||
| 148 | |||
| 149 | * 5.4.5 : used 178, allocated 515 | ||
| 150 | * 6.2 : used 118, allocated 320 | ||
| 151 | |||
| 152 | == Memory after full SOLR index == | ||
| 153 | |||
| 154 | (% style="color:orange" %)(/) Use less but keep a bit more | ||
| 155 | |||
| 156 | Delete the data/solr folder, restart, load home page then wait until nothing is happening in background. | ||
| 157 | |||
| 158 | Unit: MB | ||
| 159 | |||
| 160 | * 5.4.5 : used 273, allocated 645 | ||
| 161 | * 6.2 : used 344, allocated 503 |
