|
This isn't exactly VC++ but I'm having difficulty with the InstallShield that comes bundled with it so...
At the end of installation, in the "Finished" dialog, I want to give the user the option to display the readme file and launch the program. I know how to add those the text for those options and read the results of the user's selection but how do I actually perform those operations?
In the InstallShield script, how do you display a text file in Notepad and launch the installed software?
|
|
|
|
|
Anybody know how delete operator works over dynamically allocated 2-D array ? How is come to know about the bytes allocated ?
Vikram S
|
|
|
|
|
The memory location is contiguous.
Kuphryn
|
|
|
|
|
Yes, I meant generally arrays are allocated in continous memory locations.The issue is not that. How the delete operator comes to know about "How many bytes to be freed?"
|
|
|
|
|
I would say it deallocates memory until there is a NULL.
Kuphryn
|
|
|
|
|
hey in case of string it's quite okey. But what if i have array of intergers? and NULL =0 ?
|
|
|
|
|
can anyone help me regarding enabling dhcp automatically.
The program should change the network properties to obtain ip address dynamically.
I tried dhcpsvc and iphlpapi but i cant make it to work. anyone???
Thanks in advance.
Conrad
|
|
|
|
|
To enable DHCP:
HKEY hKey;
DWORD dwEnable = 1;
RegOpenKeyEx(HKEY_LOCAL_MACHINE, "SYSTEM\CurrentControlSet\Services\{5AE5DC8F-790F-4DA3-B5EF-492139A60ED7}\Parameters\Tcpip", 0, KEY_WRITE, &hKey);
RegSetValueEx(hKey, "EnableDHCP", 0, REG_DWORD, &dwEnable, sizeof(dwEnable));
const char szTemp[] = "0.0.0.0";
RegSetValueEx(hKey, "IPAddress", 0, REG_MULTI_SZ, (const BYTE *) szTemp, sizeof(szTemp) + 1);
RegSetValueEx(hKey, "SubnetMask", 0, REG_MULTI_SZ, (const BYTE *) szTemp, sizeof(szTemp) + 1);The bold interface will need to be changed accordingly.
"The pointy end goes in the other man." - Antonio Banderas (Zorro, 1998)
|
|
|
|
|
Thank you very much for the reply, after this how do i restart my network connection in order for the changes in the registry will take effect??
Another big thanks in advance.
|
|
|
|
|
I'm not sure. See if IpReleaseAddress() and IpRenewAddress() do anything useful.
"The pointy end goes in the other man." - Antonio Banderas (Zorro, 1998)
|
|
|
|
|
i have a program use dialog popup to get user info, but if i press ENTER instead of click on the okay button, it just act like i press the CANCEL button. How can i make it work with either click on OK button or press ENTER work the sameway?
|
|
|
|
|
In Dialog editor set Default button style to the OK button.
Robert-Antonio
"A piece of paper is an ink-lined plane.
An inclined plane is slope up.
A slow pup is a lazy dog.
Q.E.D.: A piece of paper is a lazy dog."
|
|
|
|
|
|
I have to call twice memory copy function every 33ms, now I do it with memcpy,but their cpu-cost is too high(two call cost 10ms,about %30 cpu-cost) to run program exactly .So I want to find a method to speed up mem copy .
Happy Gemini
|
|
|
|
|
memcpy is very optimized so there isn't much you can do to speed it up. However, if you ensured the memory was properly aligned, you could write some inline assembly to streamline the copy. At the very least you could measure how long it takes.
Anyone who thinks he has a better idea of what's good for people than people do is a swine.
- P.J. O'Rourke
|
|
|
|
|
i count the time every copy,it is near 5ms.
Happy Gemini
|
|
|
|
|
w_yufeng wrote:
cost 10ms,about %30 cpu-cost
 What is %30 cpu cost? During the copy the cpu usage should be 100%. If it is not then for some reason the copy is being scheduled out of the cpu. Is there any hard disk activity during the copy?
John
|
|
|
|
|
Hard Disk actiity is not done by the processor rather by DMA.
may be 30% means 10ms out of 33ms or bad math.;)
|
|
|
|
|
I was thinking that the memory copy operation was causing page faults which would explain the low cpu usage. During a page fault the cpu would have to wait for the disk. But wait a minute 10ms would probably not be enough time to read even one sector on the disk...
John
|
|
|
|
|
It depends on your computer: it can be a MMX memcpy, a 3DNow memcpy or a SSE memcpy.
BTW, what are you doing, exactly?
A pointer switch every 10ms would be way more efficient. Why can't you do it?
Or are you doing video double buffering? In this case, you can do the memcpy only on video retrace, not every 10ms. Or you could do an AGP memory transfer, using DirectX or DirectVideo.
You won't be able to speed up this so much. The newest machines have a processor of 3.2Ghz and a FSB of 2x400 Mhz, that means your processor clock is 8 times your memories' clock; in simple words: memory is becoming slower and slower...
Perl combines all the worst aspects of C and Lisp: a billion different sublanguages in one monolithic executable. It combines the power of C with the readability of PostScript. -- Jamie Zawinski
|
|
|
|
|
Daniel Turini wrote:
A pointer switch every 10ms would be way more efficient. Why can't you do it?
thanks for reply.
my app is to grab the image with ccd.but since there is lots of noise,so we must reduce it.Yn = a*Xn + b*Xn_1 + c*Yn_1 is a iterative function.I must copy current Xn to Xn_1 and save current grabed image to Xn to use next time, but its cpu usage is so high that all denoise-process cpu usage cost nearly 96% .
Happy Gemini
|
|
|
|
|
w_yufeng wrote:
I must copy current Xn to Xn_1 and save current grabed image to Xn to use next time
This is what I don't understand. Why can't you just grab the ccd image directly to Xn?
Or does the CCD provide its own memory buffers? If this is the case, you're in trouble.
If the buffers are aligned on the OS 4,096 byte boundaries, you can try to use copy-on-write to duplicate physical memory pages as logical memory pages. This would be instantaneous, although I'm not sure if this is possible under Win32, but I think that by using VirtualAlloc/VirtualProtect you may be able to do it.
Perl combines all the worst aspects of C and Lisp: a billion different sublanguages in one monolithic executable. It combines the power of C with the readability of PostScript. -- Jamie Zawinski
|
|
|
|
|
because i need save the current raw data to Xn for the next call,while the current Xn is the image that grabed last time.
you can see code below:
//get current ccd buffer
pXn = m_pAcqExBuffers->GetImageAddress(nCurBufIndex);
//m_pRawXn is the raw image data grabed in last call
memcpy((void*)m_pXn_1, (void*)m_pRawXn, 2097152);
//get ccd buffer that had been processed in last call
nPreIndex = GetBufferIndex(nCurBufIndex, -1);
pYn_1 = m_pAcqExBuffers->GetImageAddress(nPreIndex);
//save the current raw data
memcpy((void*)m_pRawXn, (void *)pXn, 2097152);
//the image process function , the pXn(it is ccd buffer) is changed in the function
DeNoisePro((WORD *)pXn, (WORD *)m_pXn_1, (WORD *)pYn_1, 1024, 1024, 0, 6, 2);
Happy Gemini
|
|
|
|
|
I understood it, but I think you did not understand me: the problem is that with the call m_pAcqExBuffers->GetImageAddress(nCurBufIndex), the responsibility for allocating memory is in the m_pAcqExBuffers object. I don't know if this class can be modified by you, but imagine if you can call something like m_pAcqExBuffers->SetImageAddress(nCurBufIndex, m_pXn_1); This way you won't need to memcpy a single byte, as the ccd will put the data right were you want.
Even if this fails, if you implement a circular buffer and preserve it between calls, you'll need only 1 memcpy per pass: you don't need to call GetBufferIndex amd GetImageAddress on the last image, as it will be on the circular buffer already.
But you'll still facing problems, as you'll be reducing a memory transfer of 400MB/second to 200MB/second. 200MB/second is still a big figure, and if you change the code as I pointed, you'll be able to avoid it completely.
Perl combines all the worst aspects of C and Lisp: a billion different sublanguages in one monolithic executable. It combines the power of C with the readability of PostScript. -- Jamie Zawinski
|
|
|
|
|
Daniel Turini wrote:
as you'll be reducing a memory transfer of 400MB/second to 200MB/second. 200MB/second is still a big figure
i dont understand .
thanks.
as you said ,the ccd buffer is circular.u can see the denoise-func below:
Yn = a*Xn + b*Yn_1 + c*Xn_1
Yn is the image that is denoised and will be showed.Xn is the raw image that ccd grab currently.Xn_1 is the raw image that was grabed last time, Yn_1 is the image that was processed last time.that is ,ccd buffer carrys raw data before process and carrys ripe data after it.we must save copy raw data to Xn_1 and Xn .so that is why I must copy data twice.
I think that there maybe not have the better method to solve the problem.Now I just pin my hope on the update of the computer hardware.
thanks a lot
Happy Gemini
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
