From Bohemia Interactive Community
|
|
Line 1: |
Line 1: |
| This article is brilliant. Thank you --[[User:Doolittle|Doolittle]] 16:57, 28 April 2010 (CEST) | | This article is brilliant. Thank you --[[User:Doolittle|Doolittle]] 16:57, 28 April 2010 (CEST) |
|
| |
| ===Trust but verify? (C)===
| |
|
| |
| It is strange that in the article aren't given the source code of the tests.
| |
| I myself decided to compare the speed ''for_from_to_do'' and ''while'':
| |
| #define __prflrStart diag_log "profiler start"; vdmj_diagProfilerTime = diag_ticktime;
| |
| #define __prflr(text) diag_log (diag_ticktime - vdmj_diagProfilerTime); diag_log text; vdmj_diagProfilerTime = diag_ticktime;
| |
|
| |
| #define __maxCount 1000000
| |
|
| |
| __prflrStart
| |
| __prflr("for begin")
| |
|
| |
| for "_i" from 1 to __maxCount do { };
| |
|
| |
| __prflr("for end")
| |
| __prflr("while begin")
| |
|
| |
| _whileCounter = 1;
| |
|
| |
| while { _whileCounter <= __maxCount } do {
| |
| _whileCounter = _whileCounter + 1;
| |
| };
| |
|
| |
| __prflr("while end")
| |
|
| |
| 1;
| |
| usage:
| |
| call compile preprocessFileLineNumbers "profiling\loops1.sqf"
| |
|
| |
| output:
| |
| "profiler start"
| |
| 0
| |
| "for begin"
| |
| 0.898438
| |
| "for end"
| |
| 0
| |
| "while begin"
| |
| 0.0351563
| |
| "while end"
| |
|
| |
| WTF?
| |
| I'd like to see in article description of the method of testing, and sqf code of the tests. Thanks
| |
|
| |
| [[User:DenV|denisko.redisko]]
| |
Revision as of 17:02, 29 May 2010
This article is brilliant. Thank you --Doolittle 16:57, 28 April 2010 (CEST)