I am using AMD gdebugger 6.2.865.0 with catalyst version 12.8 as recommended.
I tested amd gdebugger with the teapot sample and the parallax mapping sample from 2005 downloaded from the AMD page, which works.
However i tested it on two of my own OpenGL 3.3 core profile applications during which it both crashes quite early. The same happens with gdebugger 5.8 and with AMD's code XL. I tried to find heap corruption and search for other problems in my program but it was inconclusive. That being said, i want to mention that i used to debug both of my OpenGL applications with gdebugger 5.8 beforehand on another computer with an Nvidia graphics card and there it did not crash. I of course recompiled all my dependencies and the whole application on my new computer.
GPU is AMD HD Radeon 7950 with boost, if that is relevant.
I use glfw and gl3w as additional dependencies to my project. I write for Opengl 3.3 Core Profile. The program runs fine using Visual Studio 2010 in debug mode or just running the exe directly.
Here is the crash information:
If i just hit the Go (F5) button of AMD gDebugger it stops with this callstack:
ZwRaiseException - ntdll.dll
KiUserExceptionDispatcher - ntdll.dll
Which is very inconclusive and I can't do anything with that.
If i use "Draw Step (Shith+ F10)" (sic.) it will crash at this call "glClear( GL_COLOR_BUFFER_BIT );" in glfwOpenWindow() of glfw
Code of the function: http://pastebin.com/0gdVEHZt (glClear is one of the last calls there)
The crash info:
If i use Frame Step it will crash with this:
Using Step In ( F11 ) and some hundreds clicks on that, i reach this point: http://pastebin.com/j7YGZxgg
Which crashes in this call:
|_glfwWin.GetExtensionsStringEXT = (PFNWGLGETEXTENSIONSSTRINGEXTPROC)|
|wglGetProcAddress( "wglGetExtensionsStringEXT" );|
Now thats as close as i can get. Nothing seems out of order, there too, or does it?
I really need to be able to debug, track and profile my OpenGL applications, and as it has become part of my workflow to use gdebugger i would greatly appreciate support on this problem.