|
Your problem is that the main thread ends its work before the working thread gets a chance to do anything. it even does not has time for the context switch. You should use events. Check the WaitForSingleObject(), CreateEvent() and SetEvent() functions in MSDN. Generally you should wait for the event in the main thread and send it from the working thread when it's finished.
Good luck
DavidR
|
|
|
|
|
What do you mean my main thread is ending before the worker thread does anything?
I am starting the worker thread in OninitDialog() so the main thread is not ending.
Thanks for the direction, I will give those functions a whirl.
|
|
|
|
|
The start of the thread is asynchronous. IOW, you specify to start the thread and pass a fcn pointer and the call returns immediately. In the background the thread is started. Therefore, you have to decide how you want your thread to communicate back to the caller. David was referring to using events as the thread synchronization mechanism. You can find many samples and articles on doing that. Here's one I quickly grabbed from google: http://www.codersource.net/win32_waitforsingleobject.html[^] I'm sure there are several on this site as well.
If you're starting the thread from the dialog's init function, you obviously don't want to hold up that function from completing. There are several ways around this. One very easy way would be to start a timer in the dialog's OnInitDialog function where the timer function could then start the thread as per David's (and the article's) instruction.
|
|
|
|
|
Okay so answer a dumb question. Obviously I don't understand so bare with me.
How will setting things up like you say effect wheither or not my function returns a value.
Thanks.
|
|
|
|
|
There are a couple of ways to do this:
For example...
* You could have a class (CMyWorkerThread)
* CMyWorkerThread could also have a static member function that acts as the thread function (CMyWorkerThread::DoWork)
* Your client code constructs a CMyWorkerThread object and creates a thread passing the CMyWorkerThread member function pointer
* Your client code then blocks on an event that signifies that the thread has finished (e.g., by calling WaitForSingleObject)
* CMyWorkerThread::DoWork processes the work and when done it sets a member variable that represents the return value.
* CMyWorkerThread::DoWork then clears the event
* Once the CMyWorkerThread::DoWork sets the event, the client's call to WaitForSingleObject completes and the client can then access the CMyWorkerThread object's return value member variable
Another way...
* You could have a class (CMyWorkerThread)
* CMyWorkerThread has a constructor that takes a callback function that is stored as a member variable
* CMyWorkerThread has a static member function that acts as the thread function (CMyWorkerThread::DoWork)
* Your client code constructs a CMyWorkerThread object (passing the callback) and creates a thread passing the CMyWorkerThread member function pointer
* CMyWorkerThread::DoWork processes the work and when done it calls the callback function to notify the client that the work is completed and the return code
|
|
|
|
|
how can split an image into 8x8 blocks and calculate the mean,varients.. dct,fft etc.
prabhathgk
|
|
|
|
|
There is Intel mathematical library IPP which implements various functions of Image pprocessing. Check it out on the Intel web site
DavidR
|
|
|
|
|
I need to calculate the (approx) elapsed time between 2 function calls in mSec. MS recommends getting the times in SYSTEMTIMEs, converting them to FILETIMEs, copying them to ULARGE_INTEGERs and computing their difference. All this seems to work fine. But the MS docs don't indicate the resolution of the difference between the two ULARGE_INTEGERs. 
In an experiment, the time taken to execute AfxSocketInit() (i.e. the difference of the low part DWORDs) was 10000. As one would expect, introducing artificial delays caused the delta to increase. But what units am I measuring?
Thanks for any insight.
/ravi
My new year's resolution: 2048 x 1536
Home | Articles | Freeware | Music
ravib@ravib.com
|
|
|
|
|
struct _timeb tmStart, tmEnd;
_ftime (&tmStart);
foo();
_ftime (&tmEnd);
int nDeltaMSec = (tmEnd.time - tmStart.time)*1000 +
(tmEnd.millitm - tmStart.millitm);
/ravi
My new year's resolution: 2048 x 1536
Home | Articles | Freeware | Music
ravib@ravib.com
|
|
|
|
|
Hi,
I'm writing some simple image processing functions such as gaussian blur etc. I have an image class to represent the data. I then used the following to benchmark how fast it runs:
<br />
start = clock();<br />
<br />
for (int k = 0; k < 100; k++)<br />
{<br />
}<br />
<br />
<br />
finish = clock();<br />
double duration = (double)(finish - start) / CLOCKS_PER_SEC / 100;<br />
I find that my GetPixel() function is the cause of very slow performance. For example here's the code that I can benchmark:
<br />
for (int i = 0; i < test.m_nHeight; i++)<br />
{<br />
for (int j = 0; j < test.m_nWidth; j++)<br />
{<br />
int value = test.m_pData[i*test.m_nHeight + j]; <br />
}<br />
}<br />
If I access the value directly from the pointer I get duration = 0.0013 s. The other commented line of code gives me 0.012. That's 10 times slower!
I was hoping to wrap everything up nicely in a class, but it looks like that's not going to happen. Is it something I"m doing wrong, or just the way things are for function calls.
Also are there any suggestions you have on code optimization? Any articles? I would not like to start writing messy C code all wrapped up in one function for speed.
Thanks
|
|
|
|
|
Try to use inline function.
Also, check the compiler options, eg. stack checking, optimizations...
Do you build the release or debug version?
The other way is to set on assembly listing in compiler options and compare different solutions of coding.
Indeed, calling a function brings some overhead into your code. You can verify this examining mixed source and assembly listing, if compiler allows this (VC6 does allow). Generally, optimization can be a big problem and there is no perfect and only solution.
|
|
|
|
|
Yes, I tried to use inline, that didn't help. The test was done in release mode.
|
|
|
|
|
If you can step through your code, you could pin-point the costy call with the use of the @clk pseudo-register.
In your watch window, add
<br />
@clk<br />
@clk=0<br />
When you step, the value for @clk will give you how long it took.
Hope this helps
|
|
|
|
|
|
You say your GetPixel() function. What are you doing in that function? The windows GetPixel() function is very slow to begin with, so if you are calling that one then it is no wonder your function is slow. That is why DIBSections were invented, so that one could access the pixels directly as you have done instead of relying on GetPixel(). Check out the Bitmaps section here on CP for some excellent DIBSection articles.
|
|
|
|
|
My GetPixel() function is the same as the code I posted. It just dereferences the pointer. On top of that it checks if the indices are within limits and if not it returns a 0.
I rewrote my gaussian blur algorithm to use the memory directly, without GetPixel() and now my results are 10 times faster.
BTW. Thanks for the @clk tip.
|
|
|
|
|
I want to capture sound what is coming on my speaker ... not through microphone .... just the way i capture screen shot and save to disk .... is there a way to record the sound which are being played in window ... like for instance, i am listening a sound and another application is recording ... to be exact and specific that application records every sound that is played which that application is running ...
Any idea or suggestion
Any sort of help is highly appreciated
Regards,
Ibraheem Khan
|
|
|
|
|
Hi,this has been bugging me along time now and I haven´t been able to solve it yet.
I want to be able to remove a close button from an application so i can move another band to the space that was previously taken up by the close button.
I have done this successfully with a commandbar using this code:
HWND hwndCB;
TBBUTTON tbbutton;<br />
<br />
for(int i=0;i<10;i++){<br />
SendMessage(hwndCB, TB_GETBUTTON, (WPARAM)i, (LPARAM)&tbbutton);<br />
if(tbbutton.idCommand==WM_CLOSE)<br />
break;<br />
}<br />
<br />
SendMessage(hwndCB,TB_DELETEBUTTON, (WPARAM)i, (LPARAM)0 );
But unfortunately this didn´t work with the rebar control, it only removed the close button, and did not freed any space that the close button took up.
You add a close button to a rebar control using addadornments:
CommandBands_AddAdornments (hwndCB, hInst, 0, NULL);
But how in the heck do you do if you wan´t to remove it then?
There are no RemoveAdornments function and no info on this issue....
If i had the code I would have to recreate the rebar and not add any addornments to get rid of the close button.
But as i said before I don´t have the code.....
Does anyone have any idea how to solve this?
Please help, I´m going crazy...
Thanks for helping me out, Dani
|
|
|
|
|
Dialogs and Windows seem to be handled very different in Win32. Is there a way to check if a HWND is a Dialog (using DWL_DLGPROC), or a Window (using GWL_WNDPROC) ?
Pandoras Gift #44: Hope. The one that keeps you on suffering.
aber.. "Wie gesagt, der Scheiss is' Therapie"
boost your code || Fold With Us! || sighist | doxygen
|
|
|
|
|
I found this in one of my old libraries. Can't say that it's foolproof though as I don't remember using it much. (I made a function for each system class, but primarily only used AfxIsEdit, AfxIsComboBox)
BOOL AfxIsDialogBox(HWND hWnd)
{
return (AfxIsSystemClass(hWnd, _T("#32770")));
}
BOOL AfxIsSystemClass(HWND hWnd, LPCTSTR lpszSystemClassName)
{
ASSERT(::IsWindow(hWnd));
if (!::IsWindow(hWnd))
return FALSE;
ASSERT(AfxIsValidString(lpszSystemClassName));
if (!AfxIsValidString(lpszSystemClassName))
return FALSE;
TCHAR szClassName[100];
return (::GetClassName(hWnd, szClassName, 100) && _tcsicmp(szClassName, lpszSystemClassName) == 0);
}
 Pssst. You see that little light on your monitor? That's actually a government installed spy camera. Smile and wave to big brother!
|
|
|
|
|
Jack Squirrel wrote:
return (AfxIsSystemClass(hWnd, _T("#32770")));
That may not always work, as you can superclass dialogs.
--
An eye for an eye will only make the world blind.
|
|
|
|
|
How?
(Couldn't find anything in thew Dialog Box API)
Pandoras Gift #44: Hope. The one that keeps you on suffering.
aber.. "Wie gesagt, der Scheiss is' Therapie"
boost your code || Fold With Us! || sighist | doxygen
|
|
|
|
|
How to subclass, or how to detect subclassed dialogs? The answer to the first one is in the ATL library. I stumbled upon it one day as I was making my own dialog windows. The answer for the second question is: I don't know. Dialogs are tricky, that's for sure.
--
An eye for an eye will only make the world blind.
|
|
|
|
|
Jörgen Sigvardsson wrote:
That may not always work, as you can superclass dialogs.
I got an idea about using one of the dlg based functions on a non-dlg window and eventually found this system error:
ERROR_WINDOW_NOT_DIALOG - The window is not a valid dialog window
The problem is, the only dialog function I can find that returns the value is EndDialog.
if (!EndDialog(hWnd, 0))
{
if (GetLastError() == ERROR_WINDOW_NOT_DIALOG)
}
else
Pssst. You see that little light on your monitor? That's actually a government installed spy camera. Smile and wave to big brother!
|
|
|
|
|
EndDialog is not the ideal function in most cases. 
How about MapDialogRect(), or GetDlgItem()?
--
An eye for an eye will only make the world blind.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
