Page 3 of 4

Re: ATI Driver update or...

Posted: Mon Apr 12, 2010 18:56
by WinstonSmith6079
Yeah, okay. But where can I find drivers like the ones on that link of mine above, except they're v10.1? ???

Re: ATI Driver update or...

Posted: Tue Apr 27, 2010 12:55
by Sp00kyFox
found an interesting thread about this issue in the amd game forums.
don't know why but packing the gzdoom executable with upx seems to resolve the problem.
just tried it and now gzdoom works again in opengl mode wit catalyst 10.3

http://forums.amd.com/game/messageview. ... did=128067
http://upx.sourceforge.net/

Re: ATI Driver update or...

Posted: Tue Apr 27, 2010 15:11
by magmus
Sp00kyFox wrote:found an interesting thread about this issue in the amd game forums.
don't know why but packing the gzdoom executable with upx seems to resolve the problem.
just tried it and now gzdoom works again in opengl mode wit catalyst 10.3

http://forums.amd.com/game/messageview. ... did=128067
http://upx.sourceforge.net/
That works flawlessely, you should post a seperate thread about this fix!

Re: ATI Driver update or...

Posted: Tue Apr 27, 2010 17:04
by Edward-san
Executables compressed with UPX cannot use shared memory for common code. For DLL's, this means that code and data which would be using shared global memory normally would instead be using exclusive memory local to the process that loaded the library.

So one possible interpretation is that there's a bug that causes overruns in the DLL in a certain section, which, with shared memory, would try to write to memory not owned by the calling process, thereby throwing the exception. With the whole DLL loaded into the owning processes's address space, the same overrun would likely affect only memory owned by the process, which prevents the access violation. If that's the case, then there could easily be other side effects - corruption somewhere in the program, which may or may not manifest itself during gameplay.
What does it mean?

Re: ATI Driver update or...

Posted: Tue Apr 27, 2010 17:37
by Sp00kyFox
well it seems to be a nasty trick cuz it obviously destroys the possibility to load custom wads.
at least the software mode is working on amd cpus :mrgreen:

Re: ATI Driver update or...

Posted: Tue Apr 27, 2010 18:11
by Gez
Sp00kyFox wrote:well it seems to be a nasty trick cuz it obviously destroys the possibility to load custom wads.
Obviously?

Re: ATI Driver update or...

Posted: Tue Apr 27, 2010 18:26
by Sp00kyFox
sry for the confusion I meant "apparently" ;)
if packed with upx gzdoom seems to ignore every additional custom wad given by the commandline

Re: ATI Driver update or...

Posted: Tue Apr 27, 2010 19:09
by Gez
Weird.

ZDoom does not use different mechanisms to load the IWAD vs. other files, so the only explanation I see is that the command line parameters are simply lost when it's packed with UPX. This could actually make sense from a technical standpoint, since these parameters would actually be received by the unpacking process which may fail to transmit them correctly to the process of the packed executable.

Re: ATI Driver update or...

Posted: Wed Apr 28, 2010 8:23
by Rachael
And also, that seems to be a mere work-around to the problem than an actual solution.

If you pack the executable with UPX, then your crash reports become completely useless because after loading, the executable addresses are different from what they were at compile time.

Re: ATI Driver update or...

Posted: Wed Apr 28, 2010 19:44
by magmus
Sp00kyFox wrote:well it seems to be a nasty trick cuz it obviously destroys the possibility to load custom wads.
at least the software mode is working on amd cpus :mrgreen:
What on earth? I am using the latest GZdoom with AEoD v5 and it works flawlessely, even loads everything and every feature is working smooth.

EDIT:

Oh, and to add to that i am using .pk3 files and 1024cla2.wad file.

And all i did is pack the gzdoom.exe with UPX like you spotted out.

Re: ATI Driver update or...

Posted: Wed Apr 28, 2010 20:53
by Sp00kyFox
ooops.. yeah you're right, my directories were messed up.
so yup... rly seems to be a solution for all the ati users who couldn't use gzdoom any more with the new catalyst drivers.
good luck with that to all the other ati users and happy fragging ;)

:bfg:

Re: ATI Driver update or...

Posted: Wed Apr 28, 2010 23:34
by magmus
Sp00kyFox wrote:ooops.. yeah you're right, my directories were messed up.
so yup... rly seems to be a solution for all the ati users who couldn't use gzdoom any more with the new catalyst drivers.
good luck with that to all the other ati users and happy fragging ;)

:bfg:
Yeah thanks for the fix anyways!

Re: ATI Driver update or...

Posted: Thu Apr 29, 2010 3:31
by Rachael
Just remember if you crash through other means the crash reports will be useless... there won't be anything we can do about them.

Re: ATI Driver update or...

Posted: Fri Apr 30, 2010 20:04
by WinstonSmith6079
Sp00kyFox wrote:found an interesting thread about this issue in the amd game forums.
don't know why but packing the gzdoom executable with upx seems to resolve the problem.
just tried it and now gzdoom works again in opengl mode wit catalyst 10.3

http://forums.amd.com/game/messageview. ... did=128067
http://upx.sourceforge.net/
Thank you very, very much! Nice work-around--GZDoom works in full again, even with PWADs, no problemo! :D

Take note, though, that previously saved games won't work with the UPX-packed GZDoom. It will, however, create new saved games that can be loaded without issue as normal.

Thanks again! You rule! :)

Re: ATI Driver update or...

Posted: Fri Apr 30, 2010 20:40
by Sp00kyFox
WinstonSmith6079 wrote: Thank you very, very much! Nice work-around--GZDoom works in full again, even with PWADs, no problemo! :D
Take note, though, that previously saved games won't work with the UPX-packed GZDoom. It will, however, create new saved games that can be loaded without issue as normal.
Thanks again! You rule! :)
well no problem ;) I just did a little research in the amd forums, so no big deal ^^
maybe we should do an extra thread for all the ati users.
or a moderator could rename the thread so anybody can see that it includes a solution to the problem.