Calculate processing time.

[Qwert]

Hello,

I do:

Debug.WriteLine ("A: " & DateTime.Now.Ti cks.ToString)
REM Calculate a bunch of stuff. Some loops and math functions.
Debug.WriteLine ("B: " & DateTime.Now.Ti cks.ToString)

but both A and B have the exact the same value. Is this method incorrect?
When I make the function wait for 1 millisecond, A en B differ. Is there
another way to calculate the processing time? Without looping the stuff a
100 times and then devide the result by 100.

Thanks.

[Søren Reinke]

>

Debug.WriteLine ("A: " & DateTime.Now.Ti cks.ToString)
REM Calculate a bunch of stuff. Some loops and math functions.
Debug.WriteLine ("B: " & DateTime.Now.Ti cks.ToString)

but both A and B have the exact the same value. Is this method incorrect?
When I make the function wait for 1 millisecond, A en B differ. Is there
another way to calculate the processing time? Without looping the stuff a
100 times and then devide the result by 100.

I would prefer doing it 100 or even more times, to make sure you get an
accurate timing.

The first time you run a routine the CLR might do some optimizing that
actually makes it run slower the first few times.

/Søren Reinke

[Armin Zingler]

"Qwert" <no**@nosp.co m> schrieb

Hello,

I do:

Debug.WriteLine ("A: " & DateTime.Now.Ti cks.ToString)
REM Calculate a bunch of stuff. Some loops and math functions.
Debug.WriteLine ("B: " & DateTime.Now.Ti cks.ToString)

but both A and B have the exact the same value. Is this method
incorrect? When I make the function wait for 1 millisecond, A en B
differ. Is there another way to calculate the processing time?
Without looping the stuff a 100 times and then devide the result by
100.

In addition to Søren, the resolution of DateTime.Now is too low to measure
things. Use Environment.Tic kcount (returns milliseconds) instead or a high
performance counter (in VB 2005, you can use System.Diagnost ics.Stopwatch)

See also:
http://groups.google.com/group/micro...7a1a5c8d980072

(especially post #4ff ;-)

Armin

[Herfried K. Wagner [MVP]]

"Qwert" <no**@nosp.co m> schrieb:

Debug.WriteLine ("A: " & DateTime.Now.Ti cks.ToString)
REM Calculate a bunch of stuff. Some loops and math functions.
Debug.WriteLine ("B: " & DateTime.Now.Ti cks.ToString)

but both A and B have the exact the same value. Is this method incorrect?

You may want to use this timer:

<URL:http://www.mentalis.or g/soft/class.qpx?id=8>

--
M S Herfried K. Wagner
M V P <URL:http://dotnet.mvps.org/>
V B <URL:http://classicvb.org/petition/>

[Qwert]

Great, thanks all.

That 'QueryPerforman ceCounter' and 'QueryPerforman ceFrequency' code of
mentalis did the trick. None of the .NET functionality I could find was
precise enough (fastest time was 0,0003 seconds).

Don't have VB2005 yet, so I couldn't try "in VB 2005, you can use
System.Diagnost ics.Stopwatch". I did try:

Dim objProcessorTim e As New System.Diagnost ics.Performance Counter()
objProcessorTim e.CategoryName = "Processor"
objProcessorTim e.CounterName = "% Processor Time"
objProcessorTim e.InstanceName = "_Total"

but wasn't precise enough.

"Herfried K. Wagner [MVP]" <hi************ ***@gmx.at> schreef in bericht
news:OI******** ******@tk2msftn gp13.phx.gbl...

"Qwert" <no**@nosp.co m> schrieb:

Debug.WriteLine ("A: " & DateTime.Now.Ti cks.ToString)
REM Calculate a bunch of stuff. Some loops and math functions.
Debug.WriteLine ("B: " & DateTime.Now.Ti cks.ToString)

but both A and B have the exact the same value. Is this method incorrect?

You may want to use this timer:

<URL:http://www.mentalis.or g/soft/class.qpx?id=8>

--
M S Herfried K. Wagner
M V P <URL:http://dotnet.mvps.org/>
V B <URL:http://classicvb.org/petition/>