I am running a process in ~20000 images. It was crashing every time at a different point, so I decided to debug it.
After a few debugging I have realized about this:
I've just create a dummy function that equalizes an image and when it goes out of the scope it does not release all the memory.
I have tried with:
equalizedImage.release() but same result.
Any idea about what to do or if I am not managing something correctly??
(using Opencv version 2.4.9)
Thank you!!
The documentation doesn't indicate that the src and dest can be the same image.
I hate not to be able to comment for this, but I want to note a few things.
1) Windows' Task Manager memory usage indicator cannot be fully trusted for this purpose. The Process is free to keep some memory reserved for future use.
2) Please post the code that leaks so we can reproduce it properly. Possibly with tools like Valgrind the leak can be detected. At least that is better than the Task Manager.
3) Have you considered upgrading to the latest release of OpenCV to see if this behavior disappears?
4) I highly doubt that this has something to do with your crashes.
Your code is fine and should not cause any leak. The problem is somewhere else. However, there is common problem when linking openCV wrongly with your program (for example linking with MT while your program is MD or another wrong configurations) that cause a problem in releasing cv::Mat. This problem usually appears in the Debug mode and disappeared in the Release mode. so you may try to debug it in the debug mode. This problem may prevent Mat from being freed correctly.
BTW, what is the memory usage you are getting after 1000 epochs?
Related
We have a program which is proving difficult to debug. The MFC application runs fine for a few hours but over a day it will crash. Sometimes it will not throw any errors and simply exit the "dlg.DoModal()" section of our code without the user (or any obvious part of the code) closing the dialog. Other times it will crash and throw an error outside of our code and have a horrendous call stack that has nothing to do with our code, it has a lot of calls to system DLLs however.
A bit of background to our problem.
We are trying to develop a MFC bases C++ application (with a dialog). We have multiple threads and the code is rather large which makes debugging a nightmare. We have been experiencing intermittent crashes that we have been unable to locate the source of so far.
We are well past the use of breakpoints for debugging as we are pretty sure it is an issue somewhere in MFC, maybe not a bug but more a problem with the way we are using MFC.
Now we've tried simple things to help us like:
Enable all debug exceptions:
Debug -> Exceptions and then checking all the boxes so that we can
trap silly mistakes.
This proved helpful but we've now corrected all the errors it throws
within a few hours of running.
Search for memory leaks
We then tried Visual Leak Detector (which works beautifully by the
way) located here: http://vld.codeplex.com/ Our code now reports no
memory leaks so it is not an obvious memory leak issue. We have
included vld.h in the very top of our code near the entry point.
Adding Microsoft Symbol Server to obtain debug symbol files.
We then tried making our call stacks more human readable by using MS
Symbol servers shown in these tutorials:
http://social.msdn.microsoft.com/Forums/vstudio/en-US/3f1825e1-6770-48c0-91b0-12d8946ab259/2-how-do-i-configure-visual-studio-to-use-microsoft-symbol-server?forum=vsdebug and http://support.microsoft.com/kb/311503
This ultimately did nothing as it still doesn't tell you enough about any errors.
Using the Thread window to see all call stacks, and using the
Parallel Stack window
We have been using the Thread and Parallel stack windows to aid our
debugging but ultimately they have proved nothing more than pretty
pictures and fancily formatted call stacks that makes you feel good
more than anything. We have been using the tutorial here which has
been very handy
http://www.codeproject.com/Articles/79508/Mastering-Debugging-in-Visual-Studio-A-Beginn
Now for the more interesting things we've tried that do not throw errors straight away but can be detected as problems:
GDI Objects not being destroyed
This one is not obvious in VS2010 as an issue. Basically if you use a call like "CDC* pDC = lChild->GetDC();" and do not use the call "ReleaseDC(pDC);" then you have just created a GDI Object that will not be destroyed until your program terminates. VS2010 is a bit dumb in this regard and will keep creating these objects until your program crashes, and the call stack will look horrible and you will probably have no idea why it has crashed.
To find this issue, start Windows Task Manager -> Click Processes Tab -> Click the View Toolbar item -> Select Columns. Now check Handles, Threads, User Objects, and GDI Objects. Now start your program, find it's process in the list under Image Name, and watch to see if the GDI Objects column keeps growing or stabilizes.
Objects Not being destroyed
This is another not so obvious error, if you create a bitmap like this: "reinterpret_cast(LoadImage( GetModuleHandle(NULL), MAKEINTRESOURCE(IDB_BITMAPNPR), IMAGE_BITMAP, 0, 0, 0))" and assign it straight to a picture control, the bitmap will not be destroyed if you assign another bitmap to this picture control using similar code. Instead you need to assign the above to a HBITMAP variable which you then need to destroy when you are done.
This situation can also arise if you create a font or colour in a similar fashion.
Now with all that being said, we have tried all the methods above and we still can't find our issue. Sometimes our program will exit normally and we won't be given any debug info (this is usually after is has been running overnight), other times our program will lockup the PC (tested on multiple PCs), other times it will throw an error but we can't locate the culprit because it simply points to the ".DoModal()" part of our code and the rest is native windows DLLs which is useless for debugging purposes.
We suspect something is either being created and not destroyed properly but we aren't sure what and VS2010 is not telling us anything useful to point us in the right direction.
Does anyone have any ideas? How do we trap errors that aren't obvious to VS2010? Or rather how do we easily trap "GDI leaks" and the like?
Thanks in advance
Edit:
We've been using Microsoft's Application Verifier, it's found a few errors so far. To use it download it here http://www.microsoft.com/en-us/download/details.aspx?id=20028 run Application Verifier, add your .exe file in your Debug or Release directories and run the program in VS2010 as normal. VS2010 will break when Application Verifier 'sees' an error. It hasn't found anything too outrageous yet so I assume that we still have issues with our code.
I would appreciate some help concerning a good way to debug invalid memory accesses on Windows. Being a Linux developer, I am a big fan of Valgrind. Even Duma and eFence are very efficient tools. But on Windows I am totally lost (WinXP/VS2008). I tried a lot of tools, tried to link my program against duma with no luck.
Here are the symptoms for the bug I am having:
Crash in Release mode on some specific configurations (only the size of my data changes)
When it doesn't crash in Release mode, the name of the test in the report is replaced by random characters. (I am using Boost.Test)
Nothing in Debug mode, everything goes well
This definitely sounds like a invalid write/read somewhere, but debugging in Release mode is impossible, I need to reproduce the crash in Debug mode. Redefining the new operator and try to imitate the behavior of eFence/Duma was my best bet, but I am having troubles overriding it.
So far I tried:
Dr.Memory
GFlags
AppVerifier
Duma (cannot override malloc/etc..)
I'd be more than thankful if someone could give me a tip there because it's been a while since I am trying to resolve this issue.
Thanks in advance.
We have a problem for which I am looking for a direction. We have an old MFC based application, and due to various reasons, the application is crashing sometimes intermittently in some weird scenarios. And even the customers who are using our application and getting these crashes are finding difficulty in identifying the pattern of crash. So, I had a thought that, if we can have a mechanism by which we can generate a log whenever the application crashes. Like for example, the call stack or any other information in that log. I know,, we can use the crash dump in this case, but then I feel like having a log is a better option. So any help or information in this regard would be really helpful.
Thank you.
You can find a good implementation of crash reporter in the link here.
When you compile your release build, make sure that both DEBUG and /MAP are enabled. Store your binaries together with your .map files and let your customer run this version until a crash is produced. In the Event Viewer you will then find a crash log with a crash offset. Then debug step into your code (F10) and use the crash offset together with some nifty tricks and tricks to jump (set the EIP register to... well, you have to google this a bit) to the location where the crash occurred. You should now be able to find the error!
Alright, I presented this question on the MSDN forums but have yet to receive any kind of response so I figured I'd give StackOverflow a try.
I'm currently developing a DirectX application using VS2008 on Win7. I recently experienced a nasty memory corruption bug with a memory allocation class that grabbed byte aligned memory. During this bug I could still run the debug and release executables however it would crash due the instructions getting corrupted or whatever, but it would still execute for a bit until the crash.
I then stripped out the entire memory allocation class. The application runs perfectly in the IDE (release and debug builds) but I can't run any of the executables at all. They immediately crash with a non-responding/stop working error. And I don't think it is my environment because I get the same issue on another computer that wasn't having problems before either.
Dependency walker gives a "Warning: At least one delay-load dependency module was not found. Warning: At least one module has an unresolved import due to a missing export function in a delay-load dependent module." error and indicates that GPSVC.dll and IESHIMS.DLL can't be found. I've read that this can be misleading and just indicates a potential problem somewhere. And Dependency walker wasn't giving me this error the day before.
I haven't tinkered with any of the configuration or project settings or added new code. Any idea of what could be causing this behavior?
Also another note, I installed the Windows 7.1 sdk the same day. Think this could be some kind of compiler related bug?
Just in case some useful information pops up on the MSDN post, here is the link
http://social.msdn.microsoft.com/Forums/en-IE/vsdebug/thread/f692b394-8af2-4453-991c-aa6a443a9019
Thanks!
Edit -
Here is the last couple lines of Dependency Walker's profiling output
GetProcAddress(0x76CD0000 [c:\windows\syswow64\KERNEL32.DLL], "DecodePointer") called from "c:\windows\syswow64\NVWGF2UM.DLL" at address 0x6D8BAE4F and returned 0x77B59D65.
GetProcAddress(0x76CD0000 [c:\windows\syswow64\KERNEL32.DLL], "DecodePointer") called from "c:\windows\syswow64\NVWGF2UM.DLL" at address 0x6D8BAE4F and returned 0x77B59D65.
GetProcAddress(0x76CD0000 [c:\windows\syswow64\KERNEL32.DLL], "EncodePointer") called from "c:\windows\syswow64\NVWGF2UM.DLL" at address 0x6D8BAF60 and returned 0x77B60FDB.
GetProcAddress(0x76CD0000 [c:\windows\syswow64\KERNEL32.DLL], "DecodePointer") called from "c:\windows\syswow64\NVWGF2UM.DLL" at address 0x6D8BAF70 and returned 0x77B59D65.
Second chance exception 0xC0000005 (Access Violation) occurred in "c:\users\joel\desktop\DXAPP.EXE" at address 0x0110152E.
Exited "c:\users\joel\desktop\DXAPP.EXE" (process 0x27D8) with code 255 (0xFF).
Is this referring to a DLL grabbing a null pointer or to my actual instructions? Going to read up on how to use WinDbg real quick and I'll post it's output if this doesn't shed any immediate light on the issue.
Edit 2 -
Simply running the application and hitting debug to bring up Visual Studio consistently brought me to where I'm compiling my shaders. I'm assuming at the moment that the root of the problem lies around this. However, I still don't understand the change of behavior during execution between using the IDE and not.
Solution! -
I was so thrown off by the previous memory corruption bug that I didn't realize my shaders weren't in a local directory to the executables. This in turn was generating a null pointer that wasn't handled properly after calling D3DX11CompileFromFile().
Shoot, sorry I meant to post this as a comment...
I can only suggest more diagnostic attempts.
One would be to profile the app from within Depends, this will also show dynamic DLL loads and might show something new. Also it captures the debug output. It may behave differently than launching in the debugger itself and provide a clue. You don't mention actually profiling so I thought I'd suggest it in case you hadn't. Also, pay very close attention to the paths for the DLL's loaded - you might be surprised at a DLL loading from a location other than you intended.
Another suggestion is to try at attach to the stopped app after the crash (before dismissing the error dialog). See if you can get a stack trace or anything out of it.
Finally try attaching (or even launching from) WindDbg rather than the IDE. Like the Depends profile, the difference in debugger behavior and how it hooks the app may allow the crash to happen, while providing the clues you need.
Good Luck!
I am debugging an Iphone program with the simulator in xCode and I have one last issue to resolve but I need help resolving it for the following reason: when it happens the program goes into debugging mode but no errors appear (no BAD ACCESS appears) and it does not show where the code fails. Putting some variables as global helps me to see their values to start pin pointing where the bug is but before I go into this fully I would like to know what techniques/tools you guys use to debug these situations.
If it helps Im debugging the following: I merged some code into the SpeakHere demo. The code was added in the C++ modules of the program (AQRecorder.h and .mm). I seem to have pinpointed the problem code in a function I wrote.
My favourite is always to add debugging code and log it to a file. This allows me so report any and all information I need to resolve the issue if the debugger is not working properly.
I normally control the debugging code by use of a flag which I can manipulate at run time or by the command line.
If the error is (and it probably is) a memory management issue, printing log entries is really not going to help.
I would reccomend learning how to use Instruments, and use its tools to track down the memory leak when it occurs rather than waiting until the application crashes later on.