Timer test results! Timer control, timer function, gettickcount, querryperformance

Ok, i did some tests on all timer methods, and here's the results. The querryperformance counter show the accurate rate of the loop of all updating procedures. All values shows next update in milliseconds.

The test shows that gettickcount is not accurate in ms as expected

Timer control results (Using getticount counter ~53ms)
8321936
8321991
8322102
8322156
8322211
8322265
8322322
8322376
8322431
8322486
8322542
8322594
8322652
8322707
8322762
8322816
8322872
8322926
8322981
8323037
8323092
8323146
8323206
8323255
8323301
8323311
8323347
8323366
Timer function results (Normal 50-60 ms)
49549990
49550040
49550090
49550150
49550200
49550260
49550310
49550370
49550420
49550480
49550530
49550590
49550640
49550700
49550750
49550810
49550860
49550920
49550970
49551030
49551080
49551140
49551190
49551250
49551300
49551360
49553830
49553880
49553940
Gettickcount results (Normal 4 - 6ms of 1 - 8)
8871806
8871810
8871815
8871818
8871820
8871825
8871826
8871834
8871835
8871840
8871846
8871850
8871855
8871861
8871865
8871870
8871876
8871882
8871886
8871890
8871895
8871900
8871905
8871910
8871916
8871920
8871925
8871930
8871936
8871940
QuerryPerformancecounter results (Results depends on cpu speed)
9146519,1681054
9146520,19644982
9146521,34631824
9146522,40315795
9146523,42731189
9146524,4833135
9146525,5183627
9146526,63722154
9146527,67897551
9146528,69977707
9146529,7264453
9146530,74892305
9146531,98176302
9146533,01765031
9146534,04515664
9146535,07433916
9146536,11441694
9146537,13102801
9146538,190382
9146539,22710739
9146540,2596423
9146541,96097823
9146543,12258
9146544,16517206
9146545,19938316
9146547,9801874
9146549,06887477
9146550,11230493
9146554,89615984
9146556,09463786

MultiMedia Timer

Hi Kedaman, I just read this from the Visual Basic
Programmer Journal article, With the MultiMedia Timer API
function, you can create your own Timer that fully code
base and it does give you the accuratecy up to 1ms.

All you need just write some code with the following API function.

Code:

Private Declare Function timeBeginPeriod Lib "winmm.dll" (ByVal uPeriod As Long) As Long Private Declare Function timeEndPeriod Lib "winmm.dll" (ByVal uPeriod As Long) As Long Private Declare Function timeGetDevCaps Lib "winmm.dll" (lpTimeCaps As TIMECAPS, ByVal uSize As Long) As Long Private Declare Function timeGetTime Lib "winmm.dll" () As Long Private Declare Function timeKillEvent Lib "winmm.dll" (ByVal uID As Long) As Long Private Declare Function timeSetEvent Lib "winmm.dll" (ByVal uDelay As Long, ByVal uResolution As Long, ByVal lpFunction As Long, ByVal dwUser As Long, ByVal uFlags As Long) As Long

For more information about this API function, you can always refer to the MSDN library ot just download the sample code at Visual Basic Progrmmer Journal with the Locator+ number VB0004. :)

Currently I just play arround with this API function, It really cool.

Thanks! I'll do some new tests

I got the wrong results with getticount in my last tests as the tests didn't update as well because of the timer, now these are correct results: Note the small difference with gettickcount(*) and timergettime(-).

Code:

All results are in µs TIMER FUNCTION GETTICKCOUNT TIMEGETTIME QUERRYPERFORMANCE 77236610000 1601632000 1601632000 1601391142 77236610000 1601632000 1601632000 1601391245 77236610000 1601632000 1601632000 1601391306 77236610000 1601632000 1601632000 1601391365 77236610000 1601632000 1601633000 1601391434 - 77236610000 1601633000 1601633000 1601391494 * 77236610000 1601633000 1601633000 1601391554 77236610000 1601633000 1601633000 1601391613 77236610000 1601633000 1601633000 1601391670 77236610000 1601633000 1601633000 1601391728 77236610000 1601633000 1601633000 1601391787 77236610000 1601633000 1601633000 1601391845 77236610000 1601633000 1601633000 1601391903 77236610000 1601633000 1601633000 1601391961 77236610000 1601633000 1601633000 1601392019 + 77236610000 1601633000 1601633000 1601392078 77236610000 1601633000 1601633000 1601392136 77236610000 1601633000 1601633000 1601392193 77236610000 1601633000 1601633000 1601392251 77236610000 1601633000 1601633000 1601392310 77236610000 1601633000 1601633000 1601392368 77236610000 1601633000 1601634000 1601392427 - 77236610000 1601634000 1601634000 1601392484 * 77236610000 1601634000 1601634000 1601392543 77236610000 1601634000 1601634000 1601392602 77236610000 1601634000 1601634000 1601392659 77236610000 1601634000 1601634000 1601392717 77236610000 1601634000 1601634000 1601392775 77236610000 1601634000 1601634000 1601392834 77236610000 1601634000 1601634000 1601392891 77236610000 1601634000 1601634000 1601392949 77236610000 1601634000 1601634000 1601393008 + 77236610000 1601634000 1601634000 1601393065 77236610000 1601634000 1601634000 1601393124 77236610000 1601634000 1601634000 1601393181 77236610000 1601634000 1601634000 1601393239 77236610000 1601634000 1601634000 1601393296 77236610000 1601634000 1601634000 1601393355 77236610000 1601634000 1601635000 1601393415 - 77236610000 1601635000 1601635000 1601393472* 77236610000 1601635000 1601635000 1601393531 77236610000 1601635000 1601635000 1601393588 77236610000 1601635000 1601635000 1601393647 77236610000 1601635000 1601635000 1601393706 77236610000 1601635000 1601635000 1601393765 77236610000 1601635000 1601635000 1601393823 77236610000 1601635000 1601635000 1601393881 77236610000 1601635000 1601635000 1601393939+ 77236610000 1601635000 1601635000 1601394004 77236610000 1601635000 1601635000 1601394063 77236610000 1601635000 1601635000 1601394120 77236610000 1601635000 1601635000 1601394178 77236610000 1601635000 1601635000 1601394237 77236610000 1601635000 1601635000 1601394296 77236610000 1601635000 1601635000 1601394364 77236610000 1601635000 1601636000 1601394423 - 77236610000 1601636000 1601636000 1601394481 * 77236610000 1601636000 1601636000 1601394539 77236610000 1601636000 1601636000 1601394595 77236610000 1601636000 1601636000 1601394654 77236610000 1601636000 1601636000 1601394712 77236610000 1601636000 1601636000 1601394770 77236610000 1601636000 1601636000 1601394828 77236610000 1601636000 1601636000 1601394885 77236610000 1601636000 1601636000 1601394943 77236610000 1601636000 1601636000 1601395001+ 77236610000 1601636000 1601636000 1601395059 77236610000 1601636000 1601636000 1601395117 77236610000 1601636000 1601636000 1601395175 77236610000 1601636000 1601636000 1601395232 77236610000 1601636000 1601636000 1601395289 77236610000 1601636000 1601636000 1601395348 77236610000 1601636000 1601637000 1601395408 - 77236610000 1601637000 1601637000 1601395467 * 77236610000 1601637000 1601637000 1601395525 77236610000 1601637000 1601637000 1601395583 77236610000 1601637000 1601637000 1601395641 77236610000 1601637000 1601637000 1601395700 77236610000 1601637000 1601637000 1601395757 77236610000 1601637000 1601637000 1601395817 77236610000 1601637000 1601637000 1601395874 77236610000 1601637000 1601637000 1601395931 77236610000 1601637000 1601637000 1601395990 77236610000 1601637000 1601637000 1601396048 + 77236610000 1601637000 1601637000 1601396106 77236610000 1601637000 1601637000 1601396164 77236610000 1601637000 1601637000 1601396222 77236610000 1601637000 1601637000 1601396280 77236610000 1601637000 1601637000 1601396338 77236610000 1601637000 1601638000 1601396397 - 77236610000 1601638000 1601638000 1601396454 * 77236610000 1601638000 1601638000 1601396513 77236610000 1601638000 1601638000 1601396571 77236610000 1601638000 1601638000 1601396629 77236610000 1601638000 1601638000 1601396687 77236610000 1601638000 1601638000 1601396744 77236610000 1601638000 1601638000 1601396802 77236610000 1601638000 1601638000 1601396860 77236610000 1601638000 1601638000 1601396918 77236610000 1601638000 1601638000 1601396976

So in conclusion, the QuerryPerformancecounter is the best one to use???

That's right Meg, Querryperformance is the best

what exactlywas the point of all those numbers?

The point is... that all methods return different stuff, they are all converted to µs so it's easy to compare

Hmmm...If I make a cntrol usng Queryperformancecounter, than it'll be much faster than the regular VB timer?? Why didn't Microsoft think of this when they made VB?

CCRP Timer?

Hi, those were interesting results.

Know anything about the CCRP High Performance Timer Objects, available free at http://www.mvps.org/CCRP? I've been using that for a while and it seems pretty accurate.

Disagree with your test result.

Hi Kedaman, I was disagree with you test result, because i did the same Timer Function omparison as well and I found that the MultiMedia Timer API function really can go up to 1ms accuratecy.

Although the QueryPerformanceCounter and QueryPerformanceFrequency can go up to 1ms precision as well, but the refresh interval just a bit longer than the Multimedis Timer API.

Below just a summerize from the result that I get.

SetTimer API - Accuratecy up to 10ms
GetTickCount API - Accuratecy up to 10ms
QueryPerformanceCounter API - Accuratecy up to 1ms
MultiMedia Timer API - Accuratecy up to 1ms
Timer Function - Accuratecy up to 1s

So, it is better to use the MultiMedia Timer API function. If and only if your program really need that precision. Else you can go for the QueryPerformanceCounter or GetTickCount API function.

Below is those API function that I used in my test procedure:
[code]
'SetTimer API
Private Declare Function SetTimer Lib "user32" (ByVal hwnd As Long, ByVal nIDEvent As Long, ByVal uElapse As Long, ByVal lpTimerFunc As Long) As Long
Private Declare Function KillTimer Lib "user32" (ByVal hwnd As Long, ByVal nIDEvent As Long) As Long

'GetTickCount Timer
Private Declare Function GetTickCount Lib "kernel32" () As Long

'QueryPerformaceCounter Timer
Private Declare Function QueryPerformanceCounter Lib "kernel32" (lpPerformanceCount As Currency) As Long
Private Declare Function QueryPerformanceFrequency Lib "kernel32" (lpFrequency As Currency) As Long

'MultiMedia Timer
Public Declare Function timeKillEvent Lib "winmm.dll" (ByVal uId As Long) As Long
Private Declare Function timeSetEvent Lib "winmm.dll" (ByVal uDelay As Long, ByVal uResolution As Long, ByVal lpFunction As Long, ByVal dwUser As Long, ByVal uFlags As Long) As Long
Private Const TIME_ONESHOT = 0 ' program timer for single event
Private Const TIME_PERIODIC = 1 ' program for continuous periodic event
Private Const TIME_CALLBACK_FUNCTION = 0

Here is part of my test result for using various timer function and API to count for 3 second.

Code:

SetTimer GetTickCount QueryPerformance MultiMedia Timer Timer Function ------------------------------------------------------------------------------ 11071159 11067824 11065 11061875 39788 11071169 11067824 11065 11061875 39788 11071179 11067824 11065 11061875 39788 11071189 11067824 11065 11061875 39788 11071199 11067824 11065 11061886 39788 11071209 11067824 11065 11061886 39788 11071219 11067824 11065 11061886 39788 11071229 11067824 11065 11061887 39788 11071239 11067824 11065 11061887 39788 11071249 11067824 11065 11061887 39788 11071259 11067824 11065 11061887 39788 11071269 11067824 11065 11061888 39788 11071279 11067824 11065 11061888 39788 11071289 11067824 11065 11061888 39788 11071299 11067824 11065 11061888 39788 11071309 11067824 11065 11061889 39788 11071319 11067824 11065 11061889 39788 11071329 11067824 11065 11061890 39788 11071339 11067824 11065 11061890 39788 11071349 11067824 11065 11061890 39788 11071359 11067824 11065 11061890 39788 11071369 11067824 11065 11061891 39788 11071379 11067824 11065 11061891 39788 11071389 11067824 11065 11061892 39788 11071399 11067834 11065 11061892 39788 11071409 11067834 11065 11061893 39788 11071419 11067834 11065 11061893 39788 11071429 11067834 11065 11061893 39788 11071439 11067834 11065 11061894 39788 11071449 11067834 11065 11061895 39788 11071459 11067834 11065 11061896 39788 11071469 11067834 11065 11061897 39788 11071479 11067834 11065 11061898 39788 11071489 11067834 11065 11061899 39788 11071499 11067834 11065 11061900 39788 11071509 11067834 11065 11061901 39788 11071520 11067834 11065 11061902 39788 11071530 11067834 11065 11061903 39788 11071540 11067834 11065 11061904 39788 11071550 11067854 11065 11061905 39788 11071560 11067854 11065 11061906 39788 11071570 11067854 11065 11061907 39788 11071580 11067854 11065 11061908 39788 11071590 11067864 11065 11061909 39788 11071600 11067864 11065 11061910 39788 11071610 11067864 11065 11061911 39788 11071620 11067864 11065 11061912 39788 11071630 11067864 11065 11061913 39788 11071640 11067864 11065 11061914 39788 11071650 11067864 11065 11061915 39788 11071660 11067864 11065 11061916 39788 11071670 11067864 11065 11061917 39788 11071680 11067864 11065 11061918 39788 11071690 11067864 11065 11061919 39788 11071700 11067864 11065 11061920 39788 11071710 11067864 11065 11061921 39788 11071720 11067864 11065 11061922 39788 11071730 11067864 11065 11061923 39788 11071740 11067874 11065 11061924 39788 11071750 11067874 11065 11061925 39788 11071760 11067874 11065 11061926 39788 11071770 11067874 11065 11061927 39788 11071780 11067874 11065 11061928 39788 11071790 11067874 11065 11061929 39788 11071800 11067874 11065 11061930 39788 11071810 11067874 11065 11061931 39788 11071820 11067874 11065 11061932 39788 11071830 11067874 11065 11061933 39788 11071840 11067874 11065 11061934 39788 11071850 11067874 11065 11061935 39788 11071860 11067874 11065 11061946 39788 11071870 11067884 11065 11061947 39788 11071880 11067884 11065 11061948 39788 11071890 11067884 11065 11061948 39788 11071900 11067884 11065 11061949 39788 11071910 11067884 11065 11061949 39788 11071920 11067884 11065 11061950 39788 11071930 11067884 11065 11061951 39788 11071940 11067884 11065 11061951 39788 11071950 11067884 11065 11061952 39788 11071960 11067904 11065 11061953 39788 11071970 11067904 11065 11061953 39788 11071980 11067904 11065 11061954 39788 11071990 11067904 11065 11061954 39788 11072000 11067904 11065 11061955 39788 11072010 11067904 11065 11061956 39788 11072020 11067914 11065 11061957 39788 11072030 11067914 11065 11061957 39788 11072040 11067914 11065 11061958 39788 11072050 11067914 11065 11061958 39788 11072060 11067914 11065 11061959 39788 11072070 11067914 11065 11061960 39788 11072080 11067914 11065 11061961 39788 11072090 11067914 11065 11061961 39788 11072100 11067914 11065 11061962 39788 11072110 11067914 11065 11061962 39788 11072120 11067914 11065 11061963 39788 11072130 11067914 11065 11061964 39788 11072140 11067924 11065 11061965 39788 11072150 11067924 11065 11061965 39788 11072160 11067924 11065 11061966 39788 11072170 11067924 11065 11061967 39788 11072180 11067924 11065 11061968 39788 11072190 11067924 11065 11061969 39788 11072200 11067924 11065 11061970 39788 11072210 11067924 11065 11061971 39788 11072221 11067924 11065 11061972 39788 11072231 11067924 11065 11061973 39788 11072241 11067924 11065 11061974 39788 11072251 11067924 11065 11061975 39788 11072261 11067934 11065 11061976 39788 11072271 11067934 11065 11061977 39788 11072281 11067934 11065 11061978 39788 11072291 11067934 11065 11061979 39788 11072301 11067934 11065 11061980 39788 11072311 11067934 11065 11061981 39788 11072321 11067934 11065 11061982 39788 11072331 11067934 11065 11061983 39788 11072341 11067934 11065 11061984 39788 11072351 11067934 11065 11061985 39788 11072361 11067934 11065 11061986 39788 11072371 11067954 11065 11061990 39788 11072381 11067954 11065 11061991 39788 11072391 11067954 11065 11061991 39788 11072401 11067954 11065 11061992 39788 11072411 11067954 11065 11061993 39788 11072421 11067954 11065 11061994 39788 11072431 11067954 11065 11061995 39788 11072441 11067964 11065 11061995 39788 11072451 11067964 11065 11061996 39788 11072461 11067964 11065 11061997 39788 11072471 11067964 11065 11061998 39788 11072481 11067964 11065 11061999 39788 11072491 11067964 11065 11062000 39788 11072501 11067964 11065 11062000 39788 11072511 11067964 11065 11062001 39788 11072521 11067964 11065 11062002 39788 11072531 11067964 11065 11062003 39788 11072541 11067964 11065 11062004 39788 11072551 11067964 11065 11062005 39789 11072561 11067974 11065 11062006 39789 11072571 11067974 11065 11062007 39789 11072581 11067974 11065 11062008 39789 11072591 11067974 11065 11062009 39789 11072601 11067974 11065 11062010 39789 11072611 11067974 11065 11062011 39789 11072621 11067974 11065 11062013 39789 11072631 11067974 11065 11062013 39789 11072641 11067974 11065 11062014 39789 11072651 11067974 11065 11062015 39789 11072661 11067984 11065 11062016 39789 11072671 11067984 11065 11062017 39789 11072681 11067984 11065 11062018 39789 11072691 11067984 11065 11062019 39789 11072701 11067984 11065 11062020 39789 11072711 11067984 11065 11062021 39789 11072721 11067984 11065 11062022 39789 11072731 11067984 11065 11062023 39789 11072741 11067984 11065 11062024 39789 11072751 11067984 11065 11062025 39789 11072761 11067994 11065 11062026 39789 11072771 11067994 11065 11062027 39789 11072781 11067994 11065 11062028 39789 11072791 11068004 11065 11062029 39789 11072801 11068004 11065 11062030 39789 11072811 11068004 11065 11062031 39789 11072821 11068004 11065 11062032 39789 11072831 11068004 11065 11062033 39789 11072841 11068004 11065 11062034 39789 11072851 11068004 11065 11062035 39789 11072861 11068004 11065 11062037 39789 11072871 11068004 11065 11062037 39789 11072881 11068004 11065 11062039 39789 11072891 11068014 11065 11062039 39789 11072901 11068014 11065 11062041 39789 11072912 11068014 11065 11062042 39789 11072922 11068014 11065 11062043 39789 11072932 11068014 11065 11062044 39789 11072942 11068014 11065 11062045 39789 11072952 11068014 11065 11062046 39789 11072962 11068014 11065 11062047 39789 11072972 11068014 11065 11062048 39789 11072982 11068024 11065 11062049 39789 11072992 11068024 11065 11062050 39789 11073002 11068024 11065 11062051 39789 11073012 11068024 11065 11062052 39789 11073022 11068024 11065 11062053 39789 11073032 11068024 11065 11062054 39789 11073042 11068024 11065 11062055 39789 11073052 11068024 11065 11062056 39789 11073062 11068024 11065 11062057 39789 11073072 11068034 11065 11062058 39789 11073082 11068034 11065 11062059 39789 11073092 11068034 11065 11062060 39789 11073102 11068034 11065 11062061 39789 11073112 11068034 11065 11062062 39789 11073122 11068034 11065 11062063 39789 11073132 11068034 11065 11062064 39789 11073142 11068034 11065 11062065 39789 11073152 11068034 11065 11062066 39789 11073162 11068034 11066 11062067 39789 11073172 11068045 11066 11062068 39789 11073182 11068045 11066 11062069 39789 11073192 11068045 11066 11062070 39789 11073202 11068045 11066 11062071 39789 11073212 11068045 11066 11062072 39789 11073222 11068045 11066 11062074 39789 11073232 11068055 11066 11062074 39789 11073242 11068055 11066 11062076 39789 11073252 11068055 11066 11062077 39789 11073262 11068055 11066 11062078 39789 11073272 11068055 11066 11062079 39789 11073282 11068055 11066 11062080 39789 11073292 11068055 11066 11062081 39789 11073302 11068055 11066 11062083 39789 11073312 11068055 11066 11062083 39789 11073322 11068065 11066 11062085 39789 11073332 11068065 11066 11062087 39789 11073342 11068065 11066 11062088 39789 11073352 11068065 11066 11062089 39789 11073362 11068065 11066 11062091 39789 11073372 11068065 11066 11062092 39789 11073382 11068065 11066 11062093 39789 11073392 11068065 11066 11062094 39789 11073402 11068075 11066 11062095 39789 11073412 11068075 11066 11062096 39789 11073422 11068075 11066 11062098 39789 11073432 11068075 11066 11062098 39789 11073442 11068075 11066 11062100 39789 11073452 11068075 11066 11062101 39789 11073462 11068075 11066 11062102 39789 11073472 11068075 11066 11062103 39789 11073482 11068075 11066 11062107 39789 11073492 11068085 11066 11062108 39789 11073502 11068085 11066 11062110 39789 11073512 11068085 11066 11062111 39789 11073522 11068085 11066 11062112 39789 11073532 11068085 11066 11062113 39789 11073542 11068085 11066 11062115 39789 11073552 11068085 11066 11062116 39789 11073562 11068085 11066 11062117 39789 11073572 11068095 11066 11062118 39789 11073582 11068095 11066 11062119 39789 11073592 11068095 11066 11062120 39789 11073602 11068095 11066 11062122 39789 11073613 11068095 11066 11062123 39789 11073623 11068095 11066 11062124 39789 11073633 11068095 11066 11062125 39789 11073643 11068105 11066 11062127 39789 11073653 11068105 11066 11062128 39789 11073663 11068105 11066 11062129 39789 11073673 11068105 11066 11062130 39789 11073683 11068105 11066 11062132 39789 11073693 11068105 11066 11062133 39789 11073703 11068105 11066 11062135 39789 11073713 11068105 11066 11062137 39789 11073723 11068115 11066 11062138 39789 11073733 11068115 11066 11062139 39789 11073743 11068115 11066 11062141 39789 11073753 11068115 11066 11062142 39789 11073763 11068115 11066 11062143 39789 11073773 11068125 11066 11062144 39789 11073783 11068125 11066 11062146 39789 11073793 11068125 11066 11062147 39789 11073803 11068125 11066 11062148 39789 11073813 11068125 11066 11062149 39789 11073823 11068125 11066 11062151 39789 11073833 11068125 11066 11062152 39789 11073843 11068135 11066 11062153 39789 11073853 11068135 11066 11062154 39789 11073863 11068135 11066 11062156 39789 11073873 11068135 11066 11062157 39789 11073883 11068135 11066 11062159 39789 11073893 11068135 11066 11062160 39789 11073903 11068135 11066 11062161 39789 11073913 11068135 11066 11062162 39789 11073923 11068145 11066 11062164 39789 11073933 11068145 11066 11062165 39789 11073943 11068145 11066 11062167 39789 11073953 11068145 11066 11062168 39789 11073963 11068145 11066 11062169 39789 11073973 11068145 11066 11062170 39789 11073983 11068155 11066 11062172 39789 11073993 11068155 11066 11062173 39789 11074003 11068155 11066 11062175 39789 11074013 11068155 11066 11062176 39789 11074023 11068155 11066 11062177 39789 11074033 11068155 11066 11062178 39789 11074043 11068155 11066 11062180 39789 11074053 11068155 11066 11062181 39789 11074063 11068165 11066 11062183 39789 11074073 11068165 11066 11062185 39789 11074083 11068165 11066 11062187 39789 11074093 11068165 11066 11062188 39789 11074103 11068165 11066 11062190 39789 11074113 11068165 11066 11062191 39789 11074123 11068175 11066 11062192 39789 11074133 11068175 11066 11062194 39789 11074143 11068175 11066 11062195 39789 11074153 11068175 11066 11062196 39789

Kedaman,

On the weekend I worked on that performace program I mentioned in the "Inline Function" thread.

When I get a bit more done I'll post the code, so far the results are very interesting.

The Results so far look like this

FASTEST
Module Sub
Form Sub
Module Function
Foreign Form Sub
Class Sub
Form Function
Foreign Form Function
Class Function
SLOWEST

With Class functions taking 10x the time of a Module Sub.
And Form Subs taking 5 times a Module Sub
Foreign form Subs and functions are calling subs and functions from other open forms in the same app, they are are pretty close to class subs and forms in speed (not surprisingly)

I'm working on the Data Types comparison still and I'll have the code up this week, so people can suggest improvements. (with reports and graphs)

Chris, there must be something fundamentally wrong with your testing methods

Quote:

SetTimer API - Accuratecy up to 10ms
GetTickCount API - Accuratecy up to 10ms
QueryPerformanceCounter API - Accuratecy up to 1ms
MultiMedia Timer API - Accuratecy up to 1ms
Timer Function - Accuratecy up to 1s

Here are the correct results:

SetTimer API - Accuratecy up to 50-60ms (same as timer control)
GetTickCount API - Accuratecy up to 1ms
QueryPerformanceCounter API - Accuratecy up to 1µs = 0.001 ms(depending on machine)
MultiMedia Timer API - Accuratecy up to 1ms
Timer Function - Accuratecy up to 53 ms

If what you think timer returns a value in seconds, must be accurate in 1s, you're wrong, timer updates with it's decimals each ~53 second. Gettickcount is 1ms, no doubt now. Querryperformance: use CDBL to convert the division between the currency values to a floating point or you wont get a correct value. Settimer is the same as timer control.

Paul
HEhe, i know exactly why the class is slowest, but i thought that (foreign) form and class would be on the same place as they are both classes. But Module faster than function in the same form? Hmm, can you show me the results`?

MSDN Library

hi, kedaman, I don't think my test result got problem, because I just double check with MSDN library and the result does the same with mine:

Code:

HOWTO: Use QueryPerformanceCounter to Time Code ID: Q172338 -------------------------------------------------------------------------------- The information in this article applies to: Microsoft Visual Basic Control Creation, Learning, Professional, and Enterprise Editions for Windows, version 5.0 Microsoft Visual Basic Standard, Professional, and Enterprise Editions, 32-bit only, for Windows, version 4.0 Microsoft Access versions 7.0, 97 Microsoft Excel for Windows 95, version 7.0 Microsoft Excel 97 for Windows Microsoft Word for Windows -------------------------------------------------------------------------------- SUMMARY When timing code to identify performance bottlenecks, you want to use the highest resolution timer the system has to offer. This article describes how to use the QueryPerformanceCounter function to time application code. MORE INFORMATION Several timers of differing accuracy are offered by the operating system: Code: Function Units Resolution --------------------------------------------------------------------------- Now, Time, Timer seconds 1 second GetTickCount milliseconds approx. 10 ms TimeGetTime milliseconds approx. 10 ms QueryPerformanceCounter QueryPerformanceFrequency same If your system supports a high-resolution counter, you can use QueryPerformanceCounter and QueryPerformanceFrequency to do high-resolution timings. The following sample code compares the various counters: WARNING: ANY USE BY YOU OF THE CODE PROVIDED IN THIS ARTICLE IS AT YOUR OWN RISK. Microsoft provides this code "as is" without warranty of any kind, either express or implied, including but not limited to the implied warranties of merchantability and/or fitness for a particular purpose. Step-by-Step Procedures Enter the following code into a Module. If you enter it into a class, form, or report module, make the declarations Private. Option Explicit Declare Function QueryPerformanceCounter Lib "Kernel32" _ (X As Currency) As Boolean Declare Function QueryPerformanceFrequency Lib "Kernel32" _ (X As Currency) As Boolean Declare Function GetTickCount Lib "Kernel32" () As Long Declare Function timeGetTime Lib "winmm.dll" () As Long Sub Test_Timers() Dim Ctr1 As Currency, Ctr2 As Currency, Freq As Currency Dim Count1 As Long, Count2 As Long, Loops As Long ' ' Time QueryPerformanceCounter ' If QueryPerformanceCounter(Ctr1) Then QueryPerformanceCounter Ctr2 Debug.Print "Start Value: "; Format$(Ctr1, "0.0000") Debug.Print "End Value: "; Format$(Ctr2, "0.0000") QueryPerformanceFrequency Freq Debug.Print "QueryPerformanceCounter minimum resolution: 1/" & _ Freq * 10000; " sec" Debug.Print "API Overhead: "; (Ctr2 - Ctr1) / Freq; "seconds" Else Debug.Print "High-resolution counter not supported." End If ' ' Time GetTickCount ' Debug.Print Loops = 0 Count1 = GetTickCount() Do Count2 = GetTickCount() Loops = Loops + 1 Loop Until Count1 <> Count2 Debug.Print "GetTickCount minimum resolution: "; _ (Count2 - Count1); "ms" Debug.Print "Took"; Loops; "loops" ' ' Time timeGetTime ' Debug.Print Loops = 0 Count1 = timeGetTime() Do Count2 = timeGetTime() Loops = Loops + 1 Loop Until Count1 <> Count2 Debug.Print "timeGetTime minimum resolution: "; _ (Count2 - Count1); "ms" Debug.Print "Took"; Loops; "loops" End Sub Run the function from the Debug/Immediate window. Your output should appear similar to the following: Start Value: 3516284.3498 End Value: 3516284.3521 QueryPerformanceCounter minimum resolution: 1/1193182 sec API Overhead: 1.92761875388667E-05 seconds GetTickCount minimum resolution: 10 ms Took 650 loops timeGetTime minimum resolution: 10 ms Took 1565 loops Multiple statements execute before either GetTickCount or timeGetTime record a change. The actual number of loops will vary depending on the background tasks the operating system is executing. On the other hand, QueryPerformanceCounter changes value between successive API calls, indicating its usefulness in high-resolution timing. The resolution in this case is on the order of a microsecond. Because the resolution is system-dependent, there are no standard units that it measures. You have to divide the difference by the QueryPerformanceFrequency to determine the number of seconds elapsed. In the case above, the overhead for just calling the API is about 19 microseconds. This would have to be subtracted when timing other code as follows: Private Sub Time_Addition() Dim Ctr1 As Currency, Ctr2 As Currency, Freq As Currency Dim Overhead As Currency, A As Long, I As Long QueryPerformanceFrequency Freq QueryPerformanceCounter Ctr1 QueryPerformanceCounter Ctr2 Overhead = Ctr2 - Ctr1 ' determine API overhead QueryPerformanceCounter Ctr1 ' time loop For I = 1 To 100 A = A + I Next I QueryPerformanceCounter Ctr2 Debug.Print "("; Ctr1; "-"; Ctr2; "-"; Overhead; ") /"; Freq Debug.Print "100 additions took"; Debug.Print (Ctr2 - Ctr1 - Overhead) / Freq; "seconds" End Sub Sample output: ( 3630876.6256 - 3630876.6388 - 0.0013 ) / 119.3182 100 additions took 9.97333181358753E-05 seconds NOTE: Because currency variables are used, the values returned are 10000 times smaller than the actual counters. Because the calculation of seconds involves a division operation, this factor is cancelled out. REFERENCES Microsoft Developer Network; topics: timeGetTime GetTickCount QueryPerformanceCounter QueryPerformanceFrequency Additional query words: Keywords : kbprg kbVBp400 kbVBp500 kbhowto VB4WIN VBKBWinAPI vbwin Version : WINDOWS:4.0,5.0,7.0,97 Platform : WINDOWS Issue type : kbhowto Last Reviewed: October 11, 1999 ?1999 Microsoft Corporation. All rights reserved. Terms of Use. -------------------------------------------------------------------------------- Send feedback to MSDN.Look here for MSDN Online resources.

I'll send you them when I get home from work tonight (in about 4 hrs)

to be sure of the results I have added a config form so you can set the number of loops and the number of times to run that number of loops and average the data.

Usually about 2,000,000 loops run 10 times to each sub or function type (8 types) and the results averaged for each type.

A module was a long way ahead of the competition, although in real time, even a class function is pretty quich, ie, you'd have to be doing encryption, or compression or bruteforcing a decision to need the time difference.

I'll send you the project tonight, the data-types bit is only 1/4 coded but the function/subs section is done and it graphs well.

I used GetTickCount for the timing but to be honest, it wasn't really necessary.

I read documents on the net last week that had test resuld for Timer as 55-60milisecs

SetTimer API

But there was a parameter under the SetTimer API function call uElapse wich can take value upto 1ms.

I did set this value, but when I run the test, it return the tickcount to me is about 10ms different.

Code:

SetTimer The SetTimer function creates a timer with the specified time-out value. UINT_PTR SetTimer( HWND hWnd, // handle to window UINT_PTR nIDEvent, // timer identifier UINT uElapse, // time-out value TIMERPROC lpTimerFunc // timer procedure ); Parameters hWnd [in] Handle to the window to be associated with the timer. This window must be owned by the calling thread. If this parameter is NULL, no window is associated with the timer and the nIDEvent parameter is ignored. nIDEvent [in] Specifies a nonzero timer identifier. If the hWnd parameter is NULL, this parameter is ignored. If the hWnd parameter is not NULL and the window specified by hWnd already has a timer with the value nIDEvent, then the existing timer is replaced by the new timer. uElapse [in] Specifies the time-out value, in milliseconds. lpTimerFunc [in] Pointer to the function to be notified when the time-out value elapses. For more information about the function, see TimerProc. If lpTimerFunc is NULL, the system posts a WM_TIMER message to the application queue. The hwnd member of the message's MSG structure contains the value of the hWnd parameter.

Well, chris that's what's wrong, that's exactly what is wrong, my first test was done in the debug window, and the timing got totally wrong with both gettickcount and querryperformance. My new results shows that querryperformance updates as fast as your computer allow it, which is on mine about 1µs.
If it is microsoft that have taught you that i have to prove them wrong also:

Quote:

NOTE: Because currency variables are used, the values returned are 10000 times smaller than the actual counters. Because the calculation of seconds involves a division operation, this factor is cancelled out.

Code:

Private Declare Function QueryPerformanceCounter Lib "kernel32" (lpPerformanceCount As Currency) As Long Private Declare Function QueryPerformanceFrequency Lib "kernel32" (lpFrequency As Currency) As Long Dim start@, finish@, freq@ QueryPerformanceCounter start '..timeinterval QueryPerformanceCounter finish QueryPerformanceFrequency freq Interval = CDbl(finish - start) / CDbl(freq)

Try this out as It will do the division between two floating points!

Class v Module

Unless I'm really crap at reading, there was no indication as to why a class function was slower than a module function. I'd really like to know why . . .

I learn an unbelievable amount of stuff in these forums, but I think that it principally helps me learn more on my own ... springboard stylee!

Calls to objects and classes in VB comes with a certain overhead. More so in VB than in other languages there is a cost for OO implementation. Most of the ducumentation I've read suggests that classes in VB needs a bit more work, both in implemented functionality and performance.

The tests I ran show that while they're slower than module code, they're still pretty fast, (definately worth using).

So to summerise:

Modules are faster than forms which are slighly faster than classes

and

Subs are quite a bit faster than functions.

That surprised be, I expected them to be about the same.

Paul, could you do a table of tests, on arguments passed to functions and subs, and also properties, Integer, Long, Single, Double, Currency, String and Variant is enough.
That would be great.

Also, how can a form be faster than a classmodule? That's one thing i don't get.

Well, I'm outa here in a few minutes so when you get the project you can test it yourself.

maybe there is a bug in my code, or a flaw in the idea for the design

Here's my code for the second tests, if you can find a flaw in it, i'm happy to know...

Code:

Sub main() Dim a@(100, 3), n& For n = 0 To 100 DoEvents a(n, 0) = Timer * 1000000 a(n, 1) = GetTickCount * 1000 a(n, 2) = timeGetTime * 1000 QueryPerformanceCounter d@ QueryPerformanceCounter b@ QueryPerformanceFrequency c@ a(n, 3) = Int(CDbl(b@) / CDbl(c@) * 1000000) Next n For n = 0 To 100 Debug.Print a(n, 0), a(n, 1), a(n, 2), a(n, 3) Next n End Sub

Chris, how do I use the timeSetEvent and timeKillEvent? I keep crashing Vb when I use those. I need an event to be fired in each 5'th ms. As i can't do that with settimer and killtimer

Kedaman, (And anyone interested)

Here's the project zipped

http://www.ozemail.com.au/~devore/Perform.zip

Maybe we should send these results to vbapi.com and they can post up these new substitutes for the Timer Control.

QueryPerformanceCounter & QueryPerformanceFrequency API

Hi! Kedaman, I just refer back the book that I've bought (Visual Basic With Win32 API By Jason Bock (Wrox Publisher) and I found that my test program did have some problem in the QueryPerformanceCounter and QueryPerformanceFrequency API function.
According to that book, he said that the QueryperformanceCounter and QueryPerformanceFrequency can have the precision upto nanosecond which you did mention before and you're rite. :)

For my testing, I did each API function seperately instead of all together.

Here is my SetTimer and KillTimer API function call hope I did made any mistake in the test routine. ;)

Code:

Option Explicit Private T1&, T2&, dl& Private Declare Function SetTimer Lib "user32" (ByVal hwnd As Long, ByVal nIDEvent As Long, ByVal uElapse As Long, ByVal lpTimerFunc As Long) As Long Private Declare Function KillTimer Lib "user32" (ByVal hwnd As Long, ByVal nIDEvent As Long) As Long Public Sub load_SetTimer() SetTimer frmMain.hwnd, 0, 1, AddressOf TimerProc T1 = GetTickCount frmMain.cmdAction(0).Enabled = False End Sub Sub TimerProc(ByVal hwnd As Long, ByVal nIDEvent As Long, ByVal uElapse As Long, ByVal lpTimerFunc As Long) If T2 - T1 < 3000 Then T2 = GetTickCount frmMain.txtOutput(0).Text = frmMain.txtOutput(0).Text & T2 & vbCrLf Else KillTimer frmMain.hwnd, 0 End If End Sub

and this the test routine for the timeSetEvents and timeKillEvents
[code]
Public Declare Function timeKillEvent Lib "winmm.dll" (ByVal uId As Long) As Long
Private Declare Function timeSetEvent Lib "winmm.dll" (ByVal uDelay As Long, ByVal uResolution As Long, ByVal lpFunction As Long, ByVal dwUser As Long, ByVal uFlags As Long) As Long
Private Const TIME_PERIODIC = 1 ' program for continuous periodic event
Private Const TIME_CALLBACK_FUNCTION = 0
Public mTimerID&

Public Sub load_MultiMediaTimer()
mTimerID = timeSetEvent(1, 0, AddressOf OnTimer, 0, TIME_CALLBACK_FUNCTION Or TIME_PERIODIC)
T1 = timeGetTime
End Sub

Private Sub OnTimer(ByVal uId As Long, ByVal uMsg As Long, ByVal dwUser As Long, ByVal dw1 As Long, ByVal dw2 As Long)
If T2 - T1 < 3000 Then
T2 = timeGetTime
frmMain.txtOutput(3).Text = frmMain.txtOutput(3).Text & T2 & vbCrLf
Else
timeKillEvent mTimerID
End If
End Sub

Thanks Chris, i'm going to test the new events