Showing results for 
Search instead for 
Did you mean: 

Archives Discussions

Adept I

BSOD while executing SDK sample

By making a lot of modifications of SimpleGL sample (planned to base my own project on it) got repeatable BSOD (DRIVER_IRQL_NOT_LESS_OR_EQUAL in atikmdag.sys). Then, by returning to the original sample sources and making small change I've got the same result. I don't actually know is this an OpenCL runtime bug or more possible related to OpenGL implementation but it is anyway. Switch off watchdog in windows gave nothing. OS: Windows 7 x64 fully updated. Chipset: Intel x58. CPU: Intel Core i7 920 (no OC). RAM: 12GB. VCARD: AMD Juniper (Sapphire 5750) with 1024MB (no OC). Catalyst 12.6. APP SDK 2.7. All works fine with absolutely no bugs in any other applications. Just to reproduce BSOD (not to view an object) one should make a simple correction: in file SimpeGL.hpp at lines 129 and 130 replace 512 by 4096. I understand that it should be an opengl or memory overload, but I suppose that BSOD in not a proper way to show me any kind of application error. Another strange thing - I've failed to reproduce this BSOD on AMD HD6970 (with even higher mesh sizes) but will try to torment it ones more tomorrow.

5 Replies

Hi sandyandr,

So how about the original sample(nothing has been changed)? I run it, and it worked fine. Besides, I modified the SimpeGL.hpp as you said. But no BSOD. (Windows 7 x64, HD 5650, SDK2.7, driver 8.96)


Just have tried HD6970 on different system with really huge mesh sizes - just closes itself and that's all. The last check left is to put that suspicious HD5750 into this other system and try.


I think it's the driver that caused the problem. I have tested this sample(changed as you said) on 6970,7970 and 5650, the driver is 8.96, it performed fine. But on 6730, driver 8.99, it has BSOD. So I think you should update your driver to 8.96.


Actually, you seem not to be right. It's strange, but at BSOD-generating system I have the most recent drivers (12.6). And BSOD-free system has also the same drivers now - no problems there.


Wow! That's so strange. So I think there is a lot of work you will do if you want to know the reason. You should add lots of breakpoints to find out at where the program hangs. If you find the place, you will get the answer.